ASIACCS-Eval.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "42b0a1b0",
   "metadata": {},
   "source": [
    "# Evaluation of classifiers for ACM ASIACC paper\n",
    "\n",
    "128-feature vector"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "27319649",
   "metadata": {},
   "source": [
    "## Loading data and selecting features of interest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8327cd7c",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total records: 500925\n",
      "Number of used features: 143\n",
      "No. of available unused features: 36\n"
     ]
    }
   ],
   "source": [
    "import pyarrow as pa\n",
    "import pyarrow.parquet as pq\n",
    "\n",
    "#phishing = pq.read_table('floor/phishing_2307.parquet')\n",
    "phishing = pq.read_table('floor/phishing_2311.parquet')\n",
    "benign = pq.read_table('floor/benign_2310.parquet')\n",
    "\n",
    "from transformers.drop_nontrain import drop_nontrain_table as drop_nontrain\n",
    "phishing = drop_nontrain(phishing)\n",
    "benign = drop_nontrain(benign)\n",
    "\n",
    "# realign schemas (parquet files save in nonsense orders)\n",
    "benign = benign.cast(phishing.schema)\n",
    "\n",
    "# concatentate tables\n",
    "data = pa.concat_tables([phishing, benign])\n",
    "df = data.to_pandas()\n",
    "\n",
    "print(f\"Total records: {len(df)}\")\n",
    "\n",
    "from transformers.cast_timestamp import cast_timestamp\n",
    "df = cast_timestamp(df)\n",
    "\n",
    "used_features = [\n",
    "\n",
    "    # IP  ===============================\n",
    "    # Used in the original NOMS paper\n",
    "    \"ip_mean_average_rtt\",\n",
    "    \"ip_entropy\",\n",
    "    \"ip_count\",\n",
    "    #\"ip_v4_count\", \"ip_v6_count\", # Moved RDAP and RENAMED to rdap_ip\n",
    "    \n",
    "    # Newly added for ACM ASIACCS\n",
    "    \"ip_v4_ratio\",\n",
    "    #\"ip_a_aaaa_to_all_ratio\", <--- not useful for this dataset\n",
    "    \"ip_as_address_entropy\",\n",
    "    \"ip_asn_entropy\",\n",
    "    \"ip_distinct_as_count\",\n",
    "    \n",
    "    # DNS  ===============================\n",
    "    # Used in the original NOMS paper\n",
    "    \"dns_A_count\",\n",
    "    \"dns_AAAA_count\",\n",
    "    \"dns_MX_count\",\n",
    "    \"dns_NS_count\",\n",
    "    #\"dns_SOA_count\", <-- REMOVED, very low effect -------------------------------------\n",
    "    \"dns_TXT_count\",\n",
    "    \"dns_CNAME_count\",\n",
    "    \"dns_soa_primary_ns_len\",\n",
    "    \"dns_soa_primary_ns_level\", # renamed\n",
    "    \"dns_soa_primary_ns_digit_count\",\n",
    "    \"dns_soa_primary_ns_entropy\",\n",
    "    \"dns_soa_email_len\",\n",
    "    \"dns_soa_email_level\", # renamed\n",
    "    \"dns_soa_email_digit_count\",\n",
    "    \"dns_soa_email_entropy\",\n",
    "    # \"dns_soa_serial\", <-- Doesn't convey any kind of comparative information\n",
    "    \"dns_soa_refresh\",\n",
    "    \"dns_soa_retry\",\n",
    "    \"dns_soa_expire\",\n",
    "    \"dns_mx_avg_len\",\n",
    "    \"dns_mx_avg_entropy\",\n",
    "    \"dns_domain_name_in_mx\",\n",
    "    \"dns_txt_spf_exists\",\n",
    "    \"dns_txt_avg_entropy\",\n",
    "    \"dns_txt_dkim_exists\",\n",
    "    \n",
    "    \"dns_zone_level\",\n",
    "    \"dns_zone_digit_count\",\n",
    "    \"dns_zone_len\",\n",
    "    \"dns_zone_entropy\",\n",
    "    \n",
    "    # Newly added for ACM ASIACCS\n",
    "    \"dns_resolved_record_types\",\n",
    "    \"dns_has_dnskey\",\n",
    "    \"dns_dnssec_score\",\n",
    "    \"dns_ttl_avg\",\n",
    "    \"dns_ttl_stdev\",\n",
    "    \"dns_ttl_low\",\n",
    "    \"dns_ttl_mid\",\n",
    "    \"dns_ttl_distinct_count\",\n",
    "    \"dns_soa_min_ttl\", # why is this not in NOMS?!\n",
    "    \"dns_txt_external_verification_score\",\n",
    "    \"dns_txt_dmarc_exists\",\n",
    "    \n",
    "    \n",
    "    # TLS  ===============================\n",
    "    # Used in the original NOMS paper\n",
    "    \"tls_broken_chain\",\n",
    "    \"tls_expired_chain\",\n",
    "    \"tls_is_self_signed\", # <--- NEW (23-12-02)\n",
    "    \"tls_root_authority_hash\", # <--- NEW (23-12-02)\n",
    "    \"tls_leaf_authority_hash\", # <--- NEW (23-12-02)\n",
    "    \"tls_total_extension_count\",\n",
    "    \"tls_critical_extensions\",\n",
    "    \"tls_with_policies_crt_count\",\n",
    "    \"tls_percentage_crt_with_policies\",\n",
    "    \"tls_x509_anypolicy_crt_count\",\n",
    "    \"tls_iso_policy_crt_count\",\n",
    "    \"tls_joint_isoitu_policy_crt_count\",\n",
    "    # \"tls_iso_policy_oid\", <-- Basically a random value, doesn't make much sense when collected like this; policy OIDs need more research\n",
    "    # \"tls_isoitu_policy_oid\", <-- same\n",
    "    #\"tls_unknown_policy_crt_count\", <-- Abandoned, no useful values\n",
    "    \"tls_root_cert_validity_len\",\n",
    "    \"tls_leaf_cert_validity_len\",\n",
    "    \"tls_chain_len\",\n",
    "    #\"tls_root_cert_lifetime\", <---- NO!\n",
    "    #\"tls_leaf_cert_lifetime\", <---- NO!\n",
    "    \"tls_subject_count\",\n",
    "    \"tls_unique_SLD_count\",\n",
    "    \"tls_server_auth_crt_count\",\n",
    "    \"tls_client_auth_crt_count\",\n",
    "    \"tls_CA_certs_in_chain_ratio\",\n",
    "    \"tls_common_name_count\",\n",
    "    \n",
    "    # Newly added for ACM ASIACCS\n",
    "    #\"tls_has_tls\", , <-- REMOVED, very low effect -------------------------------------\n",
    "    \"tls_negotiated_version_id\",\n",
    "    \"tls_negotiated_cipher_id\",\n",
    "    \n",
    "    # LEX ===============================\n",
    "    # Used in the original NOMS paper\n",
    "    \"lex_name_len\",\n",
    "    #\"lex_digit_count\", <-- abandoned, almost the same as \"lex_sub_digit_ratio\"\n",
    "    \"lex_has_digit\",\n",
    "    \"lex_phishing_keyword_count\",\n",
    "    #\"lex_vowel_count\", <-- moved to sub\n",
    "    #\"lex_underscore_hyphen_count\", <--- same as nonalnum\n",
    "    \"lex_consecutive_chars\",\n",
    "    \"lex_tld_len\",\n",
    "    \"lex_tld_abuse_score\", # *************************************************************\n",
    "    \"lex_tld_hash\", # *************************************************************\n",
    "    \n",
    "    \"lex_sld_len\",\n",
    "    \"lex_sld_norm_entropy\", # <-- newly added feature on 23-09-29\n",
    "    # ============================= New SLD features 23-11\n",
    "    \"lex_sld_phishing_keyword_count\",\n",
    "    #\"lex_sld_digit_count\",\n",
    "    #\"lex_sld_digit_ratio\",\n",
    "    \"lex_sld_vowel_count\",\n",
    "    \"lex_sld_vowel_ratio\",\n",
    "    \"lex_sld_consonant_count\",\n",
    "    \"lex_sld_consonant_ratio\",\n",
    "    \"lex_sld_non_alphanum_count\",\n",
    "    \"lex_sld_non_alphanum_ratio\",\n",
    "    \"lex_sld_hex_count\",\n",
    "    \"lex_sld_hex_ratio\",\n",
    "    # =============================      \n",
    "    \n",
    "    \"lex_sub_count\",\n",
    "    \"lex_stld_unique_char_count\",\n",
    "    \"lex_begins_with_digit\",\n",
    "    \"lex_www_flag\",\n",
    "    \"lex_sub_max_consonant_len\",\n",
    "    \"lex_sub_norm_entropy\",\n",
    "    \"lex_sub_digit_count\",\n",
    "    \"lex_sub_digit_ratio\",\n",
    "    \"lex_sub_consonant_ratio\",\n",
    "    \"lex_sub_non_alphanum_ratio\",\n",
    "    \"lex_sub_hex_ratio\",\n",
    "\n",
    "    #\"lex_bigram_matches\",  # OLD NAME\n",
    "    #\"lex_trigram_matches\", # OLD NAME\n",
    "    \"lex_phishing_bigram_matches\",  # NEW NAME\n",
    "    \"lex_phishing_trigram_matches\", # NEW NAME\n",
    "    \"lex_phishing_tetragram_matches\", # *************************************************************\n",
    "    \"lex_phishing_pentagram_matches\", # *************************************************************\n",
    "    \"lex_sub_vowel_count\",\n",
    "    \"lex_sub_vowel_ratio\",\n",
    "    \"lex_sub_consonant_count\",\n",
    "    \"lex_sub_non_alphanum_count\",\n",
    "    \"lex_sub_hex_count\",\n",
    "    \"lex_avg_part_len\",\n",
    "    \"lex_stdev_part_lens\",\n",
    "    \"lex_longest_part_len\",\n",
    "    \"lex_shortest_sub_len\",\n",
    "    \n",
    "    # RDAP ===============================\n",
    "    # Used in the original NOMS paper\n",
    "    \"rdap_registration_period\",\n",
    "    \"rdap_domain_age\",\n",
    "    \"rdap_time_from_last_change\",\n",
    "    \"rdap_domain_active_time\",\n",
    "    \"rdap_has_dnssec\",\n",
    "    \"rdap_ip_v4_count\", \"rdap_ip_v6_count\", # <-- used for NOMS, but now prefixed with rdap_\n",
    "    \n",
    "    # Newly added for ACM ASIACCS\n",
    "    \"rdap_registrar_name_len\",\n",
    "    \"rdap_registrar_name_entropy\",\n",
    "    \"rdap_registrar_name_hash\",\n",
    "    \"rdap_registrant_name_len\",\n",
    "    \"rdap_registrant_name_entropy\",\n",
    "    \"rdap_admin_name_len\",\n",
    "    \"rdap_admin_name_entropy\",\n",
    "    \"rdap_admin_email_len\",\n",
    "    \"rdap_admin_email_entropy\",\n",
    "    \n",
    "    \"rdap_ip_shortest_v4_prefix_len\",\n",
    "    \"rdap_ip_longest_v4_prefix_len\",\n",
    "    \"rdap_ip_shortest_v6_prefix_len\",\n",
    "    \"rdap_ip_longest_v6_prefix_len\",\n",
    "    \n",
    "    \"rdap_ip_avg_admin_name_len\",\n",
    "    \"rdap_ip_avg_admin_name_entropy\",\n",
    "    \"rdap_ip_avg_admin_email_len\",\n",
    "    \"rdap_ip_avg_admin_email_entropy\",\n",
    "    \n",
    "    # GEO ===============================\n",
    "    # Used in the original NOMS paper\n",
    "    \"geo_countries_count\",\n",
    "    \"geo_continent_hash\",\n",
    "    \"geo_countries_hash\",\n",
    "    \n",
    "    # Newly added for ACM ASIACCS\n",
    "    \"geo_lat_stdev\", # <- Puvodne to pridal Adam, ale pak jsme dali pryc, ze to budem resit jinak.\n",
    "    \"geo_lon_stdev\",  #    Ale muze byt helpful pro identifikaci, jestli lokality jsou blizko sebe.    \n",
    "    #\"geo_min_lat\", # not very useful\n",
    "    #\"geo_max_lat\", # not very useful\n",
    "    #\"geo_min_lon\", # not very useful\n",
    "    #\"geo_max_lon\", # not very useful\n",
    "    \"geo_mean_lat\", # *************************************************************\n",
    "    \"geo_mean_lon\", # *************************************************************\n",
    "    #\"geo_lat_range\", # not very useful\n",
    "    #\"geo_lon_range\", # not very useful\n",
    "    \"geo_centroid_lat\", # *************************************************************\n",
    "    \"geo_centroid_lon\", # *************************************************************\n",
    "    #\"geo_estimated_area\",  # not very useful \n",
    "]\n",
    "\n",
    "\n",
    "print(\"Number of used features: \" + str(len(used_features)))\n",
    "\n",
    "# NUMBER OF AVAILABLE AND UNUSED FEATURES\n",
    "missing_columns = set(df.columns) - set(used_features) - set([\"label\"])\n",
    "print(\"No. of available unused features: \" + str(len(missing_columns)))\n",
    "\n",
    "df = df[[\"label\", *used_features]]\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "913738d6",
   "metadata": {},
   "source": [
    "## Subsampling entire dataset (for testing purposes only!)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "28b11c2b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#subsample = 0.15 # 1.0 means no subsample\n",
    "subsample = 1.0 # 1.0 means no subsample\n",
    "\n",
    "if subsample < 1.0:\n",
    "    df = df.sample(frac=subsample)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "efb8fb7e",
   "metadata": {},
   "source": [
    "## PRNG seed generator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c3e752e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "def prng_create_seeds(base_seed=None, n=1):\n",
    "    rng = np.random.default_rng()\n",
    "    if base_seed is not None:\n",
    "        rng = np.random.default_rng(base_seed)\n",
    "    return rng.integers(0, 65535, size=n)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24a01143",
   "metadata": {},
   "source": [
    "## Supressing unwanted warnings (futures etc.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "4b30d3a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore', category=FutureWarning, module='sklearn.*')\n",
    "warnings.filterwarnings('ignore', category=FutureWarning, module='xgboost.*')\n",
    "warnings.filterwarnings('ignore', message=\"Series.__getitem__ treating keys as positions is deprecated.*\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b579353e",
   "metadata": {},
   "source": [
    "## Shortening feature names (for paper purposes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2b10e171",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>label</th>\n",
       "      <th>ip_mean_average_rtt</th>\n",
       "      <th>ip_entropy</th>\n",
       "      <th>ip_count</th>\n",
       "      <th>ip_v4_ratio</th>\n",
       "      <th>ip_as_address_entropy</th>\n",
       "      <th>ip_asn_entropy</th>\n",
       "      <th>ip_distinct_as_count</th>\n",
       "      <th>dns_A_count</th>\n",
       "      <th>dns_AAAA_count</th>\n",
       "      <th>...</th>\n",
       "      <th>rdap_ip_avg_admin_email_ent</th>\n",
       "      <th>geo_countries_count</th>\n",
       "      <th>geo_continent_hash</th>\n",
       "      <th>geo_countries_hash</th>\n",
       "      <th>geo_lat_stdev</th>\n",
       "      <th>geo_lon_stdev</th>\n",
       "      <th>geo_mean_lat</th>\n",
       "      <th>geo_mean_lon</th>\n",
       "      <th>geo_centroid_lat</th>\n",
       "      <th>geo_centroid_lon</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>misp_2310:phishing</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>184</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>50.45220</td>\n",
       "      <td>30.5287</td>\n",
       "      <td>75.67830</td>\n",
       "      <td>45.79305</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>misp_2310:phishing</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>misp_2310:phishing</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>144</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>55.73860</td>\n",
       "      <td>37.6068</td>\n",
       "      <td>83.60790</td>\n",
       "      <td>56.41020</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>misp_2310:phishing</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>misp_2310:phishing</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500920</th>\n",
       "      <td>benign_2310:unknown</td>\n",
       "      <td>43.488625</td>\n",
       "      <td>4.240224</td>\n",
       "      <td>16</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.106891</td>\n",
       "      <td>1.061278</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.157633</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>187</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>37.75100</td>\n",
       "      <td>-97.8220</td>\n",
       "      <td>56.62650</td>\n",
       "      <td>-146.73300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500921</th>\n",
       "      <td>benign_2310:unknown</td>\n",
       "      <td>3.996786</td>\n",
       "      <td>4.186138</td>\n",
       "      <td>14</td>\n",
       "      <td>0.5</td>\n",
       "      <td>3.900424</td>\n",
       "      <td>0.734530</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0.190507</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>187</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>39.10270</td>\n",
       "      <td>-94.5778</td>\n",
       "      <td>58.65405</td>\n",
       "      <td>-141.86670</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500922</th>\n",
       "      <td>benign_2310:unknown</td>\n",
       "      <td>4.329750</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>12</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.149881</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>86</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>35.68930</td>\n",
       "      <td>139.6899</td>\n",
       "      <td>53.53395</td>\n",
       "      <td>209.53485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500923</th>\n",
       "      <td>benign_2310:unknown</td>\n",
       "      <td>16.280900</td>\n",
       "      <td>4.643856</td>\n",
       "      <td>10</td>\n",
       "      <td>0.5</td>\n",
       "      <td>3.043856</td>\n",
       "      <td>0.721928</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0.179189</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>65</td>\n",
       "      <td>1.19615</td>\n",
       "      <td>2.3587</td>\n",
       "      <td>51.32345</td>\n",
       "      <td>11.0482</td>\n",
       "      <td>75.19095</td>\n",
       "      <td>13.03425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500924</th>\n",
       "      <td>benign_2310:unknown</td>\n",
       "      <td>12.666600</td>\n",
       "      <td>2.918296</td>\n",
       "      <td>10</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.918296</td>\n",
       "      <td>1.760964</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.105152</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>187</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>38.57730</td>\n",
       "      <td>-92.2443</td>\n",
       "      <td>57.86595</td>\n",
       "      <td>-138.36645</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>500925 rows × 144 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                      label  ip_mean_average_rtt  ip_entropy  ip_count  \\\n",
       "0        misp_2310:phishing             0.000000   -0.000000         1   \n",
       "1        misp_2310:phishing            -1.000000   -0.000000         0   \n",
       "2        misp_2310:phishing             0.000000   -0.000000         1   \n",
       "3        misp_2310:phishing            -1.000000   -0.000000         0   \n",
       "4        misp_2310:phishing            -1.000000   -0.000000         0   \n",
       "...                     ...                  ...         ...       ...   \n",
       "500920  benign_2310:unknown            43.488625    4.240224        16   \n",
       "500921  benign_2310:unknown             3.996786    4.186138        14   \n",
       "500922  benign_2310:unknown             4.329750    4.000000        12   \n",
       "500923  benign_2310:unknown            16.280900    4.643856        10   \n",
       "500924  benign_2310:unknown            12.666600    2.918296        10   \n",
       "\n",
       "        ip_v4_ratio  ip_as_address_entropy  ip_asn_entropy  \\\n",
       "0               1.0              -0.000000       -0.000000   \n",
       "1               0.0                    NaN             NaN   \n",
       "2               1.0              -0.000000       -0.000000   \n",
       "3               0.0                    NaN             NaN   \n",
       "4               0.0                    NaN             NaN   \n",
       "...             ...                    ...             ...   \n",
       "500920          1.0               4.106891        1.061278   \n",
       "500921          0.5               3.900424        0.734530   \n",
       "500922          1.0               2.000000       -0.000000   \n",
       "500923          0.5               3.043856        0.721928   \n",
       "500924          1.0               3.918296        1.760964   \n",
       "\n",
       "        ip_distinct_as_count  dns_A_count  dns_AAAA_count  ...  \\\n",
       "0                        1.0            1               0  ...   \n",
       "1                        NaN            0               0  ...   \n",
       "2                        1.0            1               0  ...   \n",
       "3                        NaN            0               0  ...   \n",
       "4                        NaN            0               0  ...   \n",
       "...                      ...          ...             ...  ...   \n",
       "500920                   3.0            2               0  ...   \n",
       "500921                   3.0            1               1  ...   \n",
       "500922                   1.0            2               0  ...   \n",
       "500923                   2.0            1               1  ...   \n",
       "500924                   4.0            1               0  ...   \n",
       "\n",
       "        rdap_ip_avg_admin_email_ent  geo_countries_count  geo_continent_hash  \\\n",
       "0                          0.000000                    1                   3   \n",
       "1                          0.000000                    0                   0   \n",
       "2                          0.000000                    1                   0   \n",
       "3                          0.000000                    0                   0   \n",
       "4                          0.000000                    0                   0   \n",
       "...                             ...                  ...                 ...   \n",
       "500920                     0.157633                    1                   1   \n",
       "500921                     0.190507                    1                   1   \n",
       "500922                     0.149881                    1                   5   \n",
       "500923                     0.179189                    1                   3   \n",
       "500924                     0.105152                    1                   1   \n",
       "\n",
       "        geo_countries_hash  geo_lat_stdev  geo_lon_stdev  geo_mean_lat  \\\n",
       "0                      184        0.00000         0.0000      50.45220   \n",
       "1                        0        0.00000         0.0000       0.00000   \n",
       "2                      144        0.00000         0.0000      55.73860   \n",
       "3                        0        0.00000         0.0000       0.00000   \n",
       "4                        0        0.00000         0.0000       0.00000   \n",
       "...                    ...            ...            ...           ...   \n",
       "500920                 187        0.00000         0.0000      37.75100   \n",
       "500921                 187        0.00000         0.0000      39.10270   \n",
       "500922                  86        0.00000         0.0000      35.68930   \n",
       "500923                  65        1.19615         2.3587      51.32345   \n",
       "500924                 187        0.00000         0.0000      38.57730   \n",
       "\n",
       "        geo_mean_lon  geo_centroid_lat  geo_centroid_lon  \n",
       "0            30.5287          75.67830          45.79305  \n",
       "1             0.0000           0.00000           0.00000  \n",
       "2            37.6068          83.60790          56.41020  \n",
       "3             0.0000           0.00000           0.00000  \n",
       "4             0.0000           0.00000           0.00000  \n",
       "...              ...               ...               ...  \n",
       "500920      -97.8220          56.62650        -146.73300  \n",
       "500921      -94.5778          58.65405        -141.86670  \n",
       "500922      139.6899          53.53395         209.53485  \n",
       "500923       11.0482          75.19095          13.03425  \n",
       "500924      -92.2443          57.86595        -138.36645  \n",
       "\n",
       "[500925 rows x 144 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "shortening = {\n",
    "    \"lex_stld_unique_char_count\": \"lex_stld_uniq_char_cnt\",\n",
    "    \"lex_sub_max_consonant_len\": \"lex_max_conson_len\",\n",
    "    \"lex_sub_non_alphanum_count\": \"lex_sub_nonalnum_count\",\n",
    "    \"lex_sub_non_alphanum_ratio\": \"lex_sub_nonalnum_ratio\",\n",
    "    \"lex_phishing_keyword_count\": \"lex_phishing_kw_count\",\n",
    "    \"lex_sld_phishing_keyword_count\": \"lex_sld_phishing_kw_count\",\n",
    "    \n",
    "    \"dns_resolved_record_types\": \"dns_resolved_rec_types\",\n",
    "    \"dns_domain_name_in_mx\": \"dns_dn_in_mx\",\n",
    "    \"dns_txt_external_verification_score\": \"dns_txt_ext_verif_score\",\n",
    "    \"dns_zone_digit_count\": \"dns_zone_digits\",\n",
    "    \"dns_soa_primary_ns_digit_count\": \"dns_soa_prim_ns_digits\",\n",
    "    \"dns_soa_primary_ns_len\" : \"dns_soa_pri_ns_len\",\n",
    "    \"dns_soa_primary_ns_level\": \"dns_soa_pri_ns_level\",\n",
    "    \"dns_soa_primary_ns_entropy\": \"dns_soa_pri_ns_entropy\",\n",
    "    \"dns_soa_email_digit_count\": \"dns_soa_email_digits\",\n",
    "    \n",
    "    \"tls_percentage_crt_with_policies\": \"tls_percentage_with_policies\",\n",
    "    \"tls_x509_anypolicy_crt_count\": \"tls_x509_anypol_crt_count\",\n",
    "    \"tls_iso_policy_crt_count\": \"tls_iso_pol_crt_count\",\n",
    "    \"tls_joint_isoitu_policy_crt_count\": \"tls_isoitu_pol_crt_count\",\n",
    "    \n",
    "    \"rdap_ip_avg_admin_name_entropy\": \"rdap_ip_avg_admin_name_ent\",\n",
    "    \"rdap_ip_avg_admin_email_entropy\": \"rdap_ip_avg_admin_email_ent\"\n",
    "}\n",
    "\n",
    "df.rename(columns=shortening, inplace=True)\n",
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56a8b8ab",
   "metadata": {},
   "source": [
    "# NaNs will be replaced by -1 (best results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "838425da",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Handle NaNs\n",
    "df.fillna(-1, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4f736af",
   "metadata": {},
   "source": [
    "# Train-Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "3460ca64",
   "metadata": {},
   "outputs": [],
   "source": [
    "from config import Config\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "class_map = {\"benign_2310:unknown\": 0, \"misp_2310:phishing\": 1}\n",
    "\n",
    "labels = df['label'].apply(lambda x: class_map[x]) # y vector\n",
    "features = df.drop('label', axis=1).copy() # X matrix\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "  features,\n",
    "  labels,\n",
    "  test_size=0.3,\n",
    "  random_state=42,\n",
    "  shuffle=True, \n",
    "  stratify=labels\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d31198dd",
   "metadata": {},
   "source": [
    "# Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "c684da21",
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "import inspect\n",
    "import datetime\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.utils import shuffle\n",
    "import json\n",
    "\n",
    "# Filter out specific FutureWarnings from scikit-learn\n",
    "warnings.filterwarnings(\"ignore\", category=FutureWarning, module=\"sklearn.utils.validation\")\n",
    "\n",
    "shuffle_training_set = True\n",
    "\n",
    "def run_logistic_regression_multiple_seeds(seeds, X_train, y_train, X_test, y_test,\n",
    "                                           display_cm=False, display_graphs=False, verbose=False):\n",
    "    results = {\n",
    "        \"Accuracy\": [],\n",
    "        \"Precision\": [],\n",
    "        \"Recall\": [],\n",
    "        \"F1\": [],\n",
    "        \"FPR\": []\n",
    "    }\n",
    "\n",
    "    # Scale the data\n",
    "    scaler = StandardScaler()\n",
    "    x_train_scaled = scaler.fit_transform(X_train)\n",
    "    x_test_scaled = scaler.transform(X_test)\n",
    "\n",
    "    for seed in seeds:\n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"LOGISTICREGRESSION({now}): Seed {seed}: {inspect.currentframe().f_code.co_name}() Running...\")\n",
    "        \n",
    "        model_x_train = x_train_scaled\n",
    "        model_y_train = y_train\n",
    "            \n",
    "        if shuffle_training_set:\n",
    "            model_x_train, model_y_train = shuffle(\n",
    "                x_train_scaled,\n",
    "                y_train,\n",
    "                random_state=seed\n",
    "            )\n",
    "            \n",
    "            # 99% sumsample\n",
    "            model_x_train, _, model_y_train, _ = train_test_split(\n",
    "                model_x_train,\n",
    "                model_y_train,\n",
    "                test_size=0.01,\n",
    "                random_state=seed,\n",
    "                shuffle=True,\n",
    "                stratify=model_y_train\n",
    "            )\n",
    "        \n",
    "        params = {\n",
    "            \"C\": 30,  # Regularization parameter (default: 1.0)\n",
    "            \"penalty\": \"l2\",  # (default: l2)\n",
    "            \"max_iter\": 1130,  # Maximum number of iterations (default: 100)\n",
    "            \"solver\": \"liblinear\",\n",
    "            \"random_state\": seed\n",
    "        }\n",
    "        \n",
    "        model = LogisticRegression(**params)\n",
    "        model.fit(model_x_train, model_y_train)\n",
    "        y_pred = model.predict(x_test_scaled)\n",
    "        \n",
    "        # Compute metrics\n",
    "        accuracy = accuracy_score(y_test, y_pred)\n",
    "        precision = precision_score(y_test, y_pred)\n",
    "        recall = recall_score(y_test, y_pred)\n",
    "        f1 = f1_score(y_test, y_pred)\n",
    "        tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
    "        fpr = fp / (fp + tn)\n",
    "\n",
    "        results[\"Accuracy\"].append(accuracy)   # Only makes sense for balanced datasets\n",
    "        results[\"Precision\"].append(precision)\n",
    "        results[\"Recall\"].append(recall)\n",
    "        results[\"F1\"].append(f1)\n",
    "        results[\"FPR\"].append(fpr)\n",
    "        \n",
    "        now = datetime.datetime.now()\n",
    "        print(f\"LR55:\\n {now}, Seed: {seed}:\\n{results}\")\n",
    "        with open(f\"_LR55_{seed}.out\", 'a') as file:\n",
    "            file.write(\"LR55 {now}, Seed: {seed}:\\n\")\n",
    "            file.write(json.dumps(results, indent=4))\n",
    "            file.write(\"\\n\")\n",
    "        \n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"LOGISTICREGRESSION({now}): Seed {seed} done.\")\n",
    "    \n",
    "    return results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7e471654",
   "metadata": {},
   "source": [
    "# SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "40f87b76",
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "import inspect\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.utils import shuffle\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import copy\n",
    "import datetime\n",
    "import json\n",
    "\n",
    "# Filter out specific FutureWarnings\n",
    "warnings.filterwarnings(\"ignore\", category=FutureWarning, module=\"sklearn.utils.validation\")\n",
    "\n",
    "shuffle_training_set = True\n",
    "\n",
    "def run_svm_multiple_seeds(seeds, X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=False):\n",
    "    results = {\n",
    "        \"Accuracy\": [],\n",
    "        \"Precision\": [],\n",
    "        \"Recall\": [],\n",
    "        \"F1\": [],\n",
    "        \"FPR\": []\n",
    "    }\n",
    "    \n",
    "    svm_x_train = copy.deepcopy(X_train)\n",
    "    svm_x_test = copy.deepcopy(X_test)\n",
    "\n",
    "    svm_y_train = copy.deepcopy(y_train)\n",
    "    svm_y_test = copy.deepcopy(y_test)\n",
    "\n",
    "\n",
    "    subsample = 1.0  # Subsample < 1.0 for TESTING PURPOSES ONLY!\n",
    "    if subsample < 1.0:\n",
    "        svm_x_train, _, svm_y_train, _ = train_test_split(\n",
    "            svm_x_train,\n",
    "            svm_y_train,\n",
    "            test_size=1.0 - subsample,\n",
    "            random_state=42,\n",
    "            shuffle=True,\n",
    "            stratify=svm_y_train\n",
    "        )\n",
    "\n",
    "        # Check the number of samples after subsampling\n",
    "        print(f\"After Subsampling - svm_x_train: {len(svm_x_train)}, svm_y_train: {len(svm_y_train)}\")\n",
    "\n",
    "    # PREPROCESSING\n",
    "    # fill nans with 0 in X_train and X_test and y_train and y_test\n",
    "    svm_x_train = svm_x_train.fillna(0)\n",
    "    svm_x_test = svm_x_test.fillna(0)\n",
    "\n",
    "    svm_y_train = svm_y_train.fillna(0)\n",
    "    svm_y_test = svm_y_test.fillna(0)\n",
    "\n",
    "    # convert x_train to numpy array\n",
    "    svm_x_train = svm_x_train.to_numpy()\n",
    "    svm_x_test = svm_x_test.to_numpy()\n",
    "\n",
    "    svm_y_train = svm_y_train.to_numpy()\n",
    "    svm_y_test = svm_y_test.to_numpy()\n",
    "\n",
    "    # Converting False and True to 0 and 1\n",
    "    svm_x_train = np.where(svm_x_train == False, 0, svm_x_train)\n",
    "    svm_x_train = np.where(svm_x_train == True, 1, svm_x_train)\n",
    "\n",
    "    svm_x_test = np.where(svm_x_test == False, 0, svm_x_test)\n",
    "    svm_x_test = np.where(svm_x_test == True, 1, svm_x_test)\n",
    "    \n",
    "    # MinMax data scaler\n",
    "    scaler = MinMaxScaler()\n",
    "    x_train_scaled = scaler.fit_transform(svm_x_train)\n",
    "    x_test_scaled = scaler.transform(svm_x_test)\n",
    "\n",
    "    for seed in seeds:\n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"SVM({now}): Seed {seed}: {inspect.currentframe().f_code.co_name}() Running...\")\n",
    "        \n",
    "        \n",
    "        model_x_train = x_train_scaled\n",
    "        model_y_train = svm_y_train\n",
    "            \n",
    "        if shuffle_training_set:\n",
    "            model_x_train, model_y_train = shuffle(\n",
    "                x_train_scaled,\n",
    "                svm_y_train,\n",
    "                random_state=seed\n",
    "            )\n",
    "            # 99% sumsample\n",
    "            model_x_train, _, model_y_train, _ = train_test_split(\n",
    "                model_x_train,\n",
    "                model_y_train,\n",
    "                test_size=0.01,\n",
    "                random_state=seed,\n",
    "                shuffle=True,\n",
    "                stratify=model_y_train\n",
    "            )\n",
    "\n",
    "        params = {\n",
    "            \"C\": 59,\n",
    "            \"gamma\": 0.1,\n",
    "            \"kernel\": \"rbf\",  # ['rbf', 'linear', 'poly', 'sigmoid']\n",
    "            \"random_state\": seed,\n",
    "            \"random_state\": None,\n",
    "            \"shrinking\": True,\n",
    "            \"probability\": True,\n",
    "            \"verbose\": False,  # Set to False to avoid excessive output\n",
    "        }\n",
    "\n",
    "        model = SVC(**params)\n",
    "        model.fit(model_x_train, model_y_train)\n",
    "        y_pred = model.predict(x_test_scaled)\n",
    "        \n",
    "        # Compute metrics\n",
    "        accuracy = accuracy_score(y_test, y_pred)\n",
    "        precision = precision_score(y_test, y_pred)\n",
    "        recall = recall_score(y_test, y_pred)\n",
    "        f1 = f1_score(y_test, y_pred)\n",
    "        tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
    "        fpr = fp / (fp + tn)\n",
    "\n",
    "        results[\"Accuracy\"].append(accuracy)   # Only makes sense for balanced datasets\n",
    "        results[\"Precision\"].append(precision)\n",
    "        results[\"Recall\"].append(recall)\n",
    "        results[\"F1\"].append(f1)\n",
    "        results[\"FPR\"].append(fpr)\n",
    "        \n",
    "        now = datetime.datetime.now()\n",
    "        print(f\"SVM:\\n {now}, Seed: {seed}:\\n{results}\")\n",
    "        with open(\"_SVM_{seed}.out\", 'a') as file:\n",
    "            file.write(\"SVM {now}, Seed: {seed}:\\n\")\n",
    "            file.write(json.dumps(results, indent=4))\n",
    "            file.write(\"\\n\")\n",
    "        \n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"SVM({now}): Seed {seed} done.\")\n",
    "\n",
    "    return results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2cb573a9",
   "metadata": {},
   "source": [
    "# Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "98cc825f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix\n",
    "from sklearn.utils import shuffle\n",
    "import datetime\n",
    "\n",
    "shuffle_training_set = True\n",
    "\n",
    "def run_decision_tree_multiple_seeds(seeds, X_train, y_train, X_test, y_test,\n",
    "                                     display_cm=False, display_graphs=False, verbose=False):\n",
    "    results = {\n",
    "        \"Accuracy\": [],\n",
    "        \"Precision\": [],\n",
    "        \"Recall\": [],\n",
    "        \"F1\": [],\n",
    "        \"FPR\": []\n",
    "    }\n",
    "    \n",
    "    for seed in seeds:\n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"DECISIONTREE({now}): Seed {seed}: {inspect.currentframe().f_code.co_name}() Running...\")\n",
    "            \n",
    "        \n",
    "        model_x_train = X_train\n",
    "        model_y_train = y_train\n",
    "            \n",
    "        if shuffle_training_set:\n",
    "            model_x_train, model_y_train = shuffle(\n",
    "                X_train,\n",
    "                y_train,\n",
    "                random_state=seed\n",
    "            )\n",
    "            \n",
    "        params = {\n",
    "            #\"max_depth\": 30,\n",
    "            \"max_depth\": 24,\n",
    "            #\"max_leaf_nodes\": 803,\n",
    "            \"max_leaf_nodes\": 403,\n",
    "            \"min_samples_leaf\": 1,\n",
    "            \"min_samples_split\": 5,\n",
    "            \"random_state\": seed\n",
    "        }\n",
    "\n",
    "        model = DecisionTreeClassifier(**params)\n",
    "        model.fit(model_x_train, model_y_train)\n",
    "        y_pred = model.predict(X_test)\n",
    "        \n",
    "        # Compute metrics\n",
    "        accuracy = accuracy_score(y_test, y_pred)\n",
    "        precision = precision_score(y_test, y_pred)\n",
    "        recall = recall_score(y_test, y_pred)\n",
    "        f1 = f1_score(y_test, y_pred)\n",
    "        tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
    "        fpr = fp / (fp + tn)\n",
    "\n",
    "        results[\"Accuracy\"].append(accuracy)   # Only makes sense for balanced datasets\n",
    "        results[\"Precision\"].append(precision)\n",
    "        results[\"Recall\"].append(recall)\n",
    "        results[\"F1\"].append(f1)\n",
    "        results[\"FPR\"].append(fpr)\n",
    "        \n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"DECISIONTREE({now}): Seed {seed} done.\")\n",
    "    \n",
    "    return results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9fa6d98b",
   "metadata": {},
   "source": [
    "# Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "77452764",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix\n",
    "from sklearn.utils import shuffle\n",
    "import datetime\n",
    "\n",
    "shuffle_training_set = True\n",
    "\n",
    "def run_random_forest_multiple_seeds(seeds, X_train, y_train, X_test, y_test,\n",
    "                                     display_cm=False, display_graphs=False, verbose=False):\n",
    "    results = {\n",
    "        \"Accuracy\": [],\n",
    "        \"Precision\": [],\n",
    "        \"Recall\": [],\n",
    "        \"F1\": [],\n",
    "        \"FPR\": []\n",
    "    }\n",
    "    \n",
    "    for seed in seeds:\n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"RANDOMFOREST({now}): Seed {seed}: {inspect.currentframe().f_code.co_name}() Running...\")\n",
    "            \n",
    "        \n",
    "        model_x_train = X_train\n",
    "        model_y_train = y_train\n",
    "            \n",
    "        if shuffle_training_set:\n",
    "            model_x_train, model_y_train = shuffle(\n",
    "                X_train,\n",
    "                y_train,\n",
    "                random_state=seed\n",
    "            )\n",
    "            \n",
    "        # Random Forest parameters with variable seed\n",
    "        params = {\n",
    "            \"n_estimators\": 340,\n",
    "            \"criterion\": 'entropy',\n",
    "            \"max_depth\": 17,\n",
    "            \"min_samples_split\": 2,\n",
    "            \"min_samples_leaf\": 1,\n",
    "            \"max_leaf_nodes\": 450,\n",
    "            \"class_weight\": {0:1, 1:2},\n",
    "            \"random_state\": seed\n",
    "        }\n",
    "\n",
    "        # Create and train the model\n",
    "        model = RandomForestClassifier(**params)\n",
    "        model.fit(model_x_train, model_y_train)\n",
    "        y_pred = model.predict(X_test)\n",
    "        \n",
    "        # Compute metrics\n",
    "        accuracy = accuracy_score(y_test, y_pred)\n",
    "        precision = precision_score(y_test, y_pred)\n",
    "        recall = recall_score(y_test, y_pred)\n",
    "        f1 = f1_score(y_test, y_pred)\n",
    "        tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
    "        fpr = fp / (fp + tn)\n",
    "\n",
    "        results[\"Accuracy\"].append(accuracy) # Only makes sense for balanced datasets\n",
    "        results[\"Precision\"].append(precision)\n",
    "        results[\"Recall\"].append(recall)\n",
    "        results[\"F1\"].append(f1)\n",
    "        results[\"FPR\"].append(fpr)\n",
    "        \n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"RANDOMFOREST({now}): Seed {seed} done.\")\n",
    "    \n",
    "    return results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c787b515",
   "metadata": {},
   "source": [
    "# AdaBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "7585ebe6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix\n",
    "import datetime\n",
    "\n",
    "shuffle_training_set = True\n",
    "\n",
    "def run_adaboost_multiple_seeds(seeds, X_train, y_train, X_test, y_test,\n",
    "                                display_cm=False, display_graphs=False, verbose=False):\n",
    "    results = {\n",
    "        \"Accuracy\": [],\n",
    "        \"Precision\": [],\n",
    "        \"Recall\": [],\n",
    "        \"F1\": [],\n",
    "        \"FPR\": []\n",
    "    }\n",
    "    \n",
    "    for seed in seeds:\n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"ADABOOST({now}: Seed {seed}: {inspect.currentframe().f_code.co_name}() Running...\")\n",
    "            \n",
    "        \n",
    "        model_x_train = X_train\n",
    "        model_y_train = y_train\n",
    "            \n",
    "        if shuffle_training_set:\n",
    "            model_x_train, model_y_train = shuffle(\n",
    "                X_train,\n",
    "                y_train,\n",
    "                random_state=seed\n",
    "            )\n",
    "            \n",
    "        dt_params = {\n",
    "            \"max_depth\": 3,\n",
    "            \"min_samples_leaf\": 1,\n",
    "            \"min_samples_split\": 5,\n",
    "            #\"class_weight\": {0:1, 1:1},\n",
    "            \"random_state\": seed\n",
    "        }    \n",
    "        \n",
    "        # AdaBoost parameters with variable seed\n",
    "        params = {\n",
    "            \"base_estimator\": DecisionTreeClassifier(**dt_params),  #  base estimator\n",
    "            \"n_estimators\": 320,  # Number of weak learners to train iteratively\n",
    "            \"random_state\": seed\n",
    "        }\n",
    "\n",
    "        # Create and train the model\n",
    "        model = AdaBoostClassifier(**params)\n",
    "        model.fit(model_x_train, model_y_train)\n",
    "        y_pred = model.predict(X_test)\n",
    "        \n",
    "        # Compute metrics\n",
    "        accuracy = accuracy_score(y_test, y_pred)\n",
    "        precision = precision_score(y_test, y_pred, zero_division=1)\n",
    "        recall = recall_score(y_test, y_pred)\n",
    "        f1 = f1_score(y_test, y_pred)\n",
    "        tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
    "        fpr = fp / (fp + tn)\n",
    "\n",
    "        results[\"Accuracy\"].append(accuracy)   # Only makes sense for balanced datasets\n",
    "        results[\"Precision\"].append(precision)\n",
    "        results[\"Recall\"].append(recall)\n",
    "        results[\"F1\"].append(f1)\n",
    "        results[\"FPR\"].append(fpr)\n",
    "        \n",
    "        # Optionally display confusion matrix\n",
    "        if display_cm:\n",
    "            cm = confusion_matrix(y_test, y_pred, labels=model.classes_)\n",
    "            disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=model.classes_)\n",
    "            disp.plot(cmap=plt.cm.Blues)\n",
    "            plt.show()\n",
    "            \n",
    "        if display_graphs:\n",
    "            pass\n",
    "        \n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"ADABOOST({now}): Seed {seed} done.\")\n",
    "    \n",
    "    return results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8eac2d15",
   "metadata": {},
   "source": [
    "# XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "5375765b",
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "\n",
    "# Suppress specific FutureWarnings from XGBoost\n",
    "warnings.filterwarnings('ignore', category=FutureWarning, module='xgboost.*')\n",
    "# Suppress specific FutureWarnings from Sklearn\n",
    "warnings.filterwarnings('ignore', category=FutureWarning, module='sklearn.*')\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import precision_recall_fscore_support, confusion_matrix\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.utils import shuffle\n",
    "from xgboost import XGBClassifier\n",
    "import pickle\n",
    "import datetime\n",
    "\n",
    "shuffle_training_set = True\n",
    "\n",
    "def run_xgb_multiple_seeds(seeds, X_train, y_train, X_test, y_test,\n",
    "                           display_cm=False, display_graphs=False, verbose=False):\n",
    "    results = {\n",
    "        \"Accuracy\": [],\n",
    "        \"Precision\": [],\n",
    "        \"Recall\": [],\n",
    "        \"F1\": [],\n",
    "        \"FPR\": []\n",
    "    }\n",
    "\n",
    "    for seed in seeds:\n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"XGBOOST({now}): Seed {seed}: {inspect.currentframe().f_code.co_name}() Running...\")\n",
    "        \n",
    "        model_x_train = X_train\n",
    "        model_y_train = y_train\n",
    "            \n",
    "        if shuffle_training_set:\n",
    "            model_x_train, model_y_train = shuffle(\n",
    "                X_train,\n",
    "                y_train,\n",
    "                random_state=seed\n",
    "            )\n",
    "            \n",
    "        params = {\n",
    "            \"objective\": \"binary:logistic\",\n",
    "            \"tree_method\": \"gpu_hist\",\n",
    "            \"sampling_method\": \"gradient_based\",\n",
    "            \"max_depth\": 8,\n",
    "            \"eta\": 0.15,\n",
    "            \"min_child_weight\": 2.0,\n",
    "            \"subsample\": 0.6,\n",
    "            \"alpha\": 0,\n",
    "            \"gamma\": 0.1,\n",
    "            \"lambda\": 1.0,\n",
    "            \"max_delta_step\": 0,\n",
    "            \"grow_policy\": \"depthwise\",\n",
    "            \"max_bin\": 512,\n",
    "            \"n_estimators\": 550,\n",
    "            \"random_state\": seed  # Set the seed for each run\n",
    "        }\n",
    "        model = XGBClassifier(\n",
    "            **params,\n",
    "            eval_metric=[\"error\", \"logloss\", \"auc\"]\n",
    "        )\n",
    "\n",
    "        eval_set = [(model_x_train, model_y_train), (X_test, y_test)]\n",
    "        model.fit(model_x_train, model_y_train, eval_set=eval_set, verbose=False)\n",
    "\n",
    "        # Calculate metrics\n",
    "        y_pred = model.predict(X_test)\n",
    "        accuracy = accuracy_score(y_test, y_pred)\n",
    "        precision, recall, f1, _ = precision_recall_fscore_support(y_test, y_pred, average='binary')\n",
    "        tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
    "        fpr = fp / (fp + tn)\n",
    "\n",
    "        results[\"Accuracy\"].append(accuracy) # Only makes sense for balanced datasets\n",
    "        results[\"Precision\"].append(precision)\n",
    "        results[\"Recall\"].append(recall)\n",
    "        results[\"F1\"].append(f1)\n",
    "        results[\"FPR\"].append(fpr)\n",
    "\n",
    "        # Optionally print confusion matrix\n",
    "        if display_cm:\n",
    "            print(f\"Confusion Matrix for Seed {seed}:\\n\", confusion_matrix(y_test, y_pred))\n",
    "\n",
    "        # Optionally display graphs\n",
    "        if display_graphs:\n",
    "            # Extract and plot metrics from eval_set results\n",
    "            # plot evaluation\n",
    "            results = model.evals_result()\n",
    "            trees = len(results['validation_0']['error'])\n",
    "            x_axis = range(0, trees)\n",
    "\n",
    "            def plot_metric(metric):\n",
    "                import matplotlib.pyplot as plt\n",
    "                plt.rcParams[\"figure.dpi\"] = 300\n",
    "                #\n",
    "                plt.plot(x_axis, results['validation_0'][metric], label='Training set')\n",
    "                plt.plot(x_axis, results['validation_1'][metric], label='Testing set')\n",
    "                plt.legend()\n",
    "                mnames = {'error': 'Classification Error', 'logloss': 'Log Loss', 'auc': 'AUC'}\n",
    "                mname = mnames[metric]\n",
    "                plt.ylabel(mname)\n",
    "                plt.xlabel('Number of trees')\n",
    "                plt.title('XGBoost ' + mname)\n",
    "                plt.show()\n",
    "\n",
    "            plot_metric('error')\n",
    "            plot_metric('logloss')\n",
    "            plot_metric('auc')\n",
    "            \n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(\"XGBOOST(now): Seed {seed} done.\")\n",
    "\n",
    "    return results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b6b2b92",
   "metadata": {},
   "source": [
    "# LightGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "63e7c67b",
   "metadata": {},
   "outputs": [],
   "source": [
    "import lightgbm as lgb\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
    "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
    "from sklearn.utils import shuffle\n",
    "import matplotlib.pyplot as plt\n",
    "import pickle\n",
    "import datetime\n",
    "\n",
    "shuffle_training_set = True\n",
    "\n",
    "def run_lgb_multiple_seeds(seeds, X_train, y_train, X_test, y_test,\n",
    "                           display_cm=False, display_graphs=False, verbose=False):\n",
    "    results = {\n",
    "        \"Accuracy\": [],\n",
    "        \"Precision\": [],\n",
    "        \"Recall\": [],\n",
    "        \"F1\": [],\n",
    "        \"FPR\": []\n",
    "    }\n",
    "    \n",
    "    for seed in seeds:\n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"LIGHTGBM({now}) Seed {seed}: {inspect.currentframe().f_code.co_name}() Running...\")\n",
    "        \n",
    "        model_x_train = X_train\n",
    "        model_y_train = y_train\n",
    "            \n",
    "        if shuffle_training_set:\n",
    "            model_x_train, model_y_train = shuffle(\n",
    "                X_train,\n",
    "                y_train,\n",
    "                random_state=seed\n",
    "            )\n",
    "        \n",
    "        # LightGBM parameters with variable seed        \n",
    "        params = {\n",
    "            \"objective\": \"binary\",  # binary classification\n",
    "            \"boosting_type\": \"gbdt\",  # Gradient Boosting Decision Tree\n",
    "            \"n_estimators\": 897, # Number of boosted trees to fit, equivalent to number of boosting iterations\n",
    "            \"num_leaves\": 59,\n",
    "            \"max_depth\": 17,\n",
    "            \"learning_rate\": 0.15,\n",
    "            \"objective\": \"binary\",\n",
    "            \"min_child_samples\": 27,\n",
    "            \"colsample_bytree\": 0.9,\n",
    "            \"reg_lambda\": 0.11,\n",
    "            \"scale_pos_weight\": 1.1, # default: 1.0\n",
    "            \"subsample_for_bin\": 240000,\n",
    "            \"random_state\": seed,\n",
    "        }\n",
    "\n",
    "        # Create and train the model\n",
    "        model = lgb.LGBMClassifier(**params)\n",
    "        model.fit(model_x_train, model_y_train)\n",
    "        \n",
    "        # Save the model if needed\n",
    "        model_filename = f\"lightgbm_model_seed_{seed}.pickle.dat\"\n",
    "        pickle.dump(model, open(model_filename, \"wb\"))\n",
    "\n",
    "        # Predictions\n",
    "        y_pred = model.predict(X_test)\n",
    "        \n",
    "         # Calculate metrics\n",
    "        tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()\n",
    "        fpr = fp / (fp + tn)\n",
    "       \n",
    "        # Compute metrics\n",
    "        results[\"Accuracy\"].append(accuracy_score(y_test, y_pred)) # only makes sense for balanced datasets\n",
    "        results[\"Precision\"].append(precision_score(y_test, y_pred, zero_division=1))\n",
    "        results[\"Recall\"].append(recall_score(y_test, y_pred))\n",
    "        results[\"F1\"].append(f1_score(y_test, y_pred))\n",
    "        results[\"FPR\"].append(fpr)\n",
    "\n",
    "        # Optionally display confusion matrix\n",
    "        if display_cm:\n",
    "            cm = confusion_matrix(y_test, y_pred, labels=model.classes_)\n",
    "            disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=model.classes_)\n",
    "            disp.plot(cmap=plt.cm.Blues)\n",
    "            plt.show()\n",
    "            \n",
    "        if display_graphs:\n",
    "            pass\n",
    "        \n",
    "        if verbose:\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"LIGHTGBM({now}): Seed {seed} done.\")\n",
    "\n",
    "    return results\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a585fae2",
   "metadata": {},
   "source": [
    "# Run evaluations"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95e766ba",
   "metadata": {},
   "source": [
    "## Initialize results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "54759041",
   "metadata": {},
   "outputs": [],
   "source": [
    "replace_results_if_nonempty = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "35e617c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    if isinstance(my_variable, dict):\n",
    "        if len(results) > 0 and replace_results_if_nonempty:\n",
    "            results = dict()\n",
    "    else:\n",
    "        results = dict()\n",
    "except NameError:\n",
    "    results = dict()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33038e24",
   "metadata": {},
   "source": [
    "## Run"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fb7c743",
   "metadata": {},
   "source": [
    "### Serial run"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "05f0f6ad",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start Time: 2023-12-06 00:53:43.255819\n",
      "LOGISTICREGRESSION(2023-12-06 00:53:45.141144): Seed 5849: run_logistic_regression_multiple_seeds() Running...\n",
      "LR3:\n",
      " 2023-12-06 01:06:01.577985, Seed: 5849:\n",
      "{'Accuracy': [0.9639202012270592], 'Precision': [0.9065768194070081], 'Recall': [0.8201014337267142], 'F1': [0.8611736993035641], 'FPR': [0.013354190426286102]}\n",
      "LOGISTICREGRESSION(2023-12-06 01:06:01.578119): Seed 5849 done.\n",
      "LOGISTICREGRESSION(2023-12-06 01:06:01.578130): Seed 50721: run_logistic_regression_multiple_seeds() Running...\n",
      "LR3:\n",
      " 2023-12-06 01:22:45.821118, Seed: 50721:\n",
      "{'Accuracy': [0.9639202012270592, 0.9638470035534143], 'Precision': [0.9065768194070081, 0.9065213873455957], 'Recall': [0.8201014337267142, 0.8195650053642837], 'F1': [0.8611736993035641, 0.8608528620822129], 'FPR': [0.013354190426286102, 0.013354190426286102]}\n",
      "LOGISTICREGRESSION(2023-12-06 01:22:45.821254): Seed 50721 done.\n",
      "LOGISTICREGRESSION(2023-12-06 01:22:45.821264): Seed 42897: run_logistic_regression_multiple_seeds() Running...\n",
      "LR3:\n",
      " 2023-12-06 01:41:00.534123, Seed: 42897:\n",
      "{'Accuracy': [0.9639202012270592, 0.9638470035534143, 0.9638470035534143], 'Precision': [0.9065768194070081, 0.9065213873455957, 0.9063460397908017], 'Recall': [0.8201014337267142, 0.8195650053642837, 0.8197600702233493], 'F1': [0.8611736993035641, 0.8608528620822129, 0.8608813663482959], 'FPR': [0.013354190426286102, 0.013354190426286102, 0.013385013716364086]}\n",
      "LOGISTICREGRESSION(2023-12-06 01:41:00.534265): Seed 42897 done.\n",
      "LOGISTICREGRESSION(2023-12-06 01:41:00.534275): Seed 28761: run_logistic_regression_multiple_seeds() Running...\n",
      "LR3:\n",
      " 2023-12-06 02:07:09.424537, Seed: 28761:\n",
      "{'Accuracy': [0.9639202012270592, 0.9638470035534143, 0.9638470035534143, 0.9638536578873821], 'Precision': [0.9065768194070081, 0.9065213873455957, 0.9063460397908017, 0.9064825800884478], 'Recall': [0.8201014337267142, 0.8195650053642837, 0.8197600702233493, 0.8196625377938165], 'F1': [0.8611736993035641, 0.8608528620822129, 0.8608813663482959, 0.8608891620569556], 'FPR': [0.013354190426286102, 0.013354190426286102, 0.013385013716364086, 0.013361896248805597]}\n",
      "LOGISTICREGRESSION(2023-12-06 02:07:09.424680): Seed 28761 done.\n",
      "LOGISTICREGRESSION(2023-12-06 02:07:09.424689): Seed 28377: run_logistic_regression_multiple_seeds() Running...\n",
      "LR3:\n",
      " 2023-12-06 02:21:14.582503, Seed: 28377:\n",
      "{'Accuracy': [0.9639202012270592, 0.9638470035534143, 0.9638470035534143, 0.9638536578873821, 0.963693953872157], 'Precision': [0.9065768194070081, 0.9065213873455957, 0.9063460397908017, 0.9064825800884478, 0.907108850898074], 'Recall': [0.8201014337267142, 0.8195650053642837, 0.8197600702233493, 0.8196625377938165, 0.8176631229883936], 'F1': [0.8611736993035641, 0.8608528620822129, 0.8608813663482959, 0.8608891620569556, 0.860066683765068], 'FPR': [0.013354190426286102, 0.013354190426286102, 0.013385013716364086, 0.013361896248805597, 0.01323089726597417]}\n",
      "LOGISTICREGRESSION(2023-12-06 02:21:14.582641): Seed 28377 done.\n",
      "End Time: 2023-12-06 02:21:14.584207\n",
      "Time Taken: 1:27:31.328388\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fitcrack/.local/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "import datetime\n",
    "import time\n",
    "\n",
    "# Print the current date and time\n",
    "start_time = datetime.datetime.now()\n",
    "print(f\"Start Time: {start_time}\")\n",
    "\n",
    "x = 42\n",
    "ROUNDS = 5\n",
    "\n",
    "# Run the function with the desired number of seeds and store the results in the dictionary\n",
    "results['Logistic Regression'] = run_logistic_regression_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['SVM'] = run_svm_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['DecisionTree'] = run_decision_tree_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['RandomForest'] = run_random_forest_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['AdaBoost'] = run_adaboost_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['XGBoost'] = run_xgb_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['LightGBM'] = run_lgb_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "\n",
    "# Print end time and duration\n",
    "end_time = datetime.datetime.now()\n",
    "print(f\"End Time: {end_time}\")\n",
    "time_taken = end_time - start_time\n",
    "print(f\"Time Taken: {time_taken}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "769d6d00",
   "metadata": {},
   "source": [
    "### Parallel run\n",
    "This is fast but may have **ultra-high** processor and memory demands.\n",
    "\n",
    "Set the **n_workers** accordingly to your machine's capabilities! Otherwise, the threads may crash due to insufficient memory, etc."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "877bb45e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import concurrent.futures\n",
    "\n",
    "def train_models_in_parallel(tasks, n_workers, X_train, y_train, X_test, y_test):\n",
    "    print(\"******** PARALLEL MODEL TRAINING & VALIDATION ********\")\n",
    "    # Print the current date and time\n",
    "    start_time = datetime.datetime.now()\n",
    "    print(f\"Start Time: {start_time}\")\n",
    "    \n",
    "    # Dictionary to store results\n",
    "    results = {}\n",
    "\n",
    "    with concurrent.futures.ProcessPoolExecutor(max_workers=n_workers) as executor:\n",
    "        # Launching the model training tasks\n",
    "        future_to_model = {}\n",
    "        for name, (task, seeds) in tasks.items():\n",
    "            now = datetime.datetime.now()\n",
    "            print(f\"{now}: Starting task: {name}\")  # Verbose: indicate task start\n",
    "            future = executor.submit(task, seeds, X_train, y_train, X_test, y_test, False, False, True)\n",
    "            future_to_model[future] = name\n",
    "\n",
    "        # Collecting results as they complete\n",
    "        for future in concurrent.futures.as_completed(future_to_model):\n",
    "            model_name = future_to_model[future]\n",
    "            now = datetime.datetime.now()\n",
    "            try:\n",
    "                results[model_name] = future.result()\n",
    "                print(f\"{now}: Completed task: {model_name}\")  # Verbose: indicate task completion\n",
    "            except Exception as exc:\n",
    "                print(f'{now}: {model_name} generated an exception: {exc}')\n",
    "                \n",
    "    # Print end time and duration\n",
    "    end_time = datetime.datetime.now()\n",
    "    print(f\"End Time: {end_time}\")\n",
    "    time_taken = end_time - start_time\n",
    "    print(f\"Time Taken: {time_taken}\")\n",
    "    print(\"******************** JOB DONE ************************\")\n",
    "\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ba3535aa",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "******** PARALLEL MODEL TRAINING & VALIDATION ********\n",
      "Start Time: 2023-12-07 20:25:28.957226\n",
      "2023-12-07 20:25:28.957458: Starting task: Logistic Regression\n",
      "2023-12-07 20:25:28.992645: Starting task: Logistic Regression4_b\n",
      "2023-12-07 20:25:28.992890: Starting task: Logistic Regression4_c\n",
      "2023-12-07 20:25:28.992911: Starting task: Logistic Regression4_d\n",
      "2023-12-07 20:25:28.992943: Starting task: Logistic Regression4_e\n",
      "LOGISTICREGRESSION(2023-12-07 20:25:32.370583): Seed 5849: run_logistic_regression_multiple_seeds() Running...\n",
      "LOGISTICREGRESSION(2023-12-07 20:25:33.388246): Seed 1031: run_logistic_regression_multiple_seeds() Running...\n",
      "LOGISTICREGRESSION(2023-12-07 20:25:34.253688): Seed 48400: run_logistic_regression_multiple_seeds() Running...\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[29], line 22\u001b[0m, in \u001b[0;36mtrain_models_in_parallel\u001b[0;34m(tasks, n_workers, X_train, y_train, X_test, y_test)\u001b[0m\n\u001b[1;32m     21\u001b[0m \u001b[38;5;66;03m# Collecting results as they complete\u001b[39;00m\n\u001b[0;32m---> 22\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m future \u001b[38;5;129;01min\u001b[39;00m concurrent\u001b[38;5;241m.\u001b[39mfutures\u001b[38;5;241m.\u001b[39mas_completed(future_to_model):\n\u001b[1;32m     23\u001b[0m     model_name \u001b[38;5;241m=\u001b[39m future_to_model[future]\n",
      "File \u001b[0;32m/usr/lib/python3.11/concurrent/futures/_base.py:243\u001b[0m, in \u001b[0;36mas_completed\u001b[0;34m(fs, timeout)\u001b[0m\n\u001b[1;32m    239\u001b[0m         \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mTimeoutError\u001b[39;00m(\n\u001b[1;32m    240\u001b[0m                 \u001b[38;5;124m'\u001b[39m\u001b[38;5;132;01m%d\u001b[39;00m\u001b[38;5;124m (of \u001b[39m\u001b[38;5;132;01m%d\u001b[39;00m\u001b[38;5;124m) futures unfinished\u001b[39m\u001b[38;5;124m'\u001b[39m \u001b[38;5;241m%\u001b[39m (\n\u001b[1;32m    241\u001b[0m                 \u001b[38;5;28mlen\u001b[39m(pending), total_futures))\n\u001b[0;32m--> 243\u001b[0m \u001b[43mwaiter\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mevent\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mwait\u001b[49m\u001b[43m(\u001b[49m\u001b[43mwait_timeout\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    245\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m waiter\u001b[38;5;241m.\u001b[39mlock:\n",
      "File \u001b[0;32m/usr/lib/python3.11/threading.py:622\u001b[0m, in \u001b[0;36mEvent.wait\u001b[0;34m(self, timeout)\u001b[0m\n\u001b[1;32m    621\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m signaled:\n\u001b[0;32m--> 622\u001b[0m     signaled \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_cond\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mwait\u001b[49m\u001b[43m(\u001b[49m\u001b[43mtimeout\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    623\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m signaled\n",
      "File \u001b[0;32m/usr/lib/python3.11/threading.py:320\u001b[0m, in \u001b[0;36mCondition.wait\u001b[0;34m(self, timeout)\u001b[0m\n\u001b[1;32m    319\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m timeout \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[0;32m--> 320\u001b[0m     \u001b[43mwaiter\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43macquire\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    321\u001b[0m     gotit \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mTrue\u001b[39;00m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: ",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[30], line 20\u001b[0m\n\u001b[1;32m      7\u001b[0m n_workers \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m3\u001b[39m  \u001b[38;5;66;03m# 0 uses all available processors\u001b[39;00m\n\u001b[1;32m      9\u001b[0m batch1 \u001b[38;5;241m=\u001b[39m {\n\u001b[1;32m     10\u001b[0m     \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mLogistic Regression\u001b[39m\u001b[38;5;124m'\u001b[39m: (run_logistic_regression_multiple_seeds, prng_create_seeds(n\u001b[38;5;241m=\u001b[39mROUNDS,base_seed\u001b[38;5;241m=\u001b[39mbase_seed)),\n\u001b[1;32m     11\u001b[0m     \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mLogistic Regression4_b\u001b[39m\u001b[38;5;124m'\u001b[39m: (run_logistic_regression_multiple_seeds, prng_create_seeds(n\u001b[38;5;241m=\u001b[39mROUNDS,base_seed\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m142\u001b[39m)),\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     17\u001b[0m     \u001b[38;5;66;03m#'RandomForest': (run_random_forest_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=base_seed)),\u001b[39;00m\n\u001b[1;32m     18\u001b[0m }\n\u001b[0;32m---> 20\u001b[0m results1 \u001b[38;5;241m=\u001b[39m \u001b[43mtrain_models_in_parallel\u001b[49m\u001b[43m(\u001b[49m\u001b[43mbatch1\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mn_workers\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mX_train\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43my_train\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mX_test\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43my_test\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     21\u001b[0m results\u001b[38;5;241m.\u001b[39mupdate(results1)\n\u001b[1;32m     23\u001b[0m \u001b[38;5;66;03m# XGboost, and LightGBM are multi-threaded by design (running them in threads may cause crashes)\u001b[39;00m\n\u001b[1;32m     24\u001b[0m \u001b[38;5;66;03m#results['AdaBoost'] = run_adaboost_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\u001b[39;00m\n\u001b[1;32m     25\u001b[0m \u001b[38;5;66;03m#results['XGBoost'] = run_xgb_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\u001b[39;00m\n\u001b[1;32m     26\u001b[0m \u001b[38;5;66;03m#results['LightGBM'] = run_lgb_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\u001b[39;00m\n",
      "Cell \u001b[0;32mIn[29], line 12\u001b[0m, in \u001b[0;36mtrain_models_in_parallel\u001b[0;34m(tasks, n_workers, X_train, y_train, X_test, y_test)\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[38;5;66;03m# Dictionary to store results\u001b[39;00m\n\u001b[1;32m     10\u001b[0m results \u001b[38;5;241m=\u001b[39m {}\n\u001b[0;32m---> 12\u001b[0m \u001b[43m\u001b[49m\u001b[38;5;28;43;01mwith\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[43mconcurrent\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfutures\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mProcessPoolExecutor\u001b[49m\u001b[43m(\u001b[49m\u001b[43mmax_workers\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mn_workers\u001b[49m\u001b[43m)\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43;01mas\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[43mexecutor\u001b[49m\u001b[43m:\u001b[49m\n\u001b[1;32m     13\u001b[0m \u001b[43m    \u001b[49m\u001b[38;5;66;43;03m# Launching the model training tasks\u001b[39;49;00m\n\u001b[1;32m     14\u001b[0m \u001b[43m    \u001b[49m\u001b[43mfuture_to_model\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m \u001b[49m\u001b[43m{\u001b[49m\u001b[43m}\u001b[49m\n\u001b[1;32m     15\u001b[0m \u001b[43m    \u001b[49m\u001b[38;5;28;43;01mfor\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[43mname\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43m(\u001b[49m\u001b[43mtask\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mseeds\u001b[49m\u001b[43m)\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;129;43;01min\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[43mtasks\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mitems\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m:\u001b[49m\n",
      "File \u001b[0;32m/usr/lib/python3.11/concurrent/futures/_base.py:647\u001b[0m, in \u001b[0;36mExecutor.__exit__\u001b[0;34m(self, exc_type, exc_val, exc_tb)\u001b[0m\n\u001b[1;32m    646\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m__exit__\u001b[39m(\u001b[38;5;28mself\u001b[39m, exc_type, exc_val, exc_tb):\n\u001b[0;32m--> 647\u001b[0m     \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mshutdown\u001b[49m\u001b[43m(\u001b[49m\u001b[43mwait\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m)\u001b[49m\n\u001b[1;32m    648\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;01mFalse\u001b[39;00m\n",
      "File \u001b[0;32m/usr/lib/python3.11/concurrent/futures/process.py:834\u001b[0m, in \u001b[0;36mProcessPoolExecutor.shutdown\u001b[0;34m(self, wait, cancel_futures)\u001b[0m\n\u001b[1;32m    831\u001b[0m         \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_executor_manager_thread_wakeup\u001b[38;5;241m.\u001b[39mwakeup()\n\u001b[1;32m    833\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_executor_manager_thread \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m wait:\n\u001b[0;32m--> 834\u001b[0m     \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_executor_manager_thread\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mjoin\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    835\u001b[0m \u001b[38;5;66;03m# To reduce the risk of opening too many files, remove references to\u001b[39;00m\n\u001b[1;32m    836\u001b[0m \u001b[38;5;66;03m# objects that use file descriptors.\u001b[39;00m\n\u001b[1;32m    837\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_executor_manager_thread \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n",
      "File \u001b[0;32m/usr/lib/python3.11/threading.py:1112\u001b[0m, in \u001b[0;36mThread.join\u001b[0;34m(self, timeout)\u001b[0m\n\u001b[1;32m   1109\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mRuntimeError\u001b[39;00m(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mcannot join current thread\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m   1111\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m timeout \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[0;32m-> 1112\u001b[0m     \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_wait_for_tstate_lock\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1113\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m   1114\u001b[0m     \u001b[38;5;66;03m# the behavior of a negative timeout isn't documented, but\u001b[39;00m\n\u001b[1;32m   1115\u001b[0m     \u001b[38;5;66;03m# historically .join(timeout=x) for x<0 has acted as if timeout=0\u001b[39;00m\n\u001b[1;32m   1116\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_wait_for_tstate_lock(timeout\u001b[38;5;241m=\u001b[39m\u001b[38;5;28mmax\u001b[39m(timeout, \u001b[38;5;241m0\u001b[39m))\n",
      "File \u001b[0;32m/usr/lib/python3.11/threading.py:1132\u001b[0m, in \u001b[0;36mThread._wait_for_tstate_lock\u001b[0;34m(self, block, timeout)\u001b[0m\n\u001b[1;32m   1129\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m\n\u001b[1;32m   1131\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m-> 1132\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[43mlock\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43macquire\u001b[49m\u001b[43m(\u001b[49m\u001b[43mblock\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtimeout\u001b[49m\u001b[43m)\u001b[49m:\n\u001b[1;32m   1133\u001b[0m         lock\u001b[38;5;241m.\u001b[39mrelease()\n\u001b[1;32m   1134\u001b[0m         \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_stop()\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LOGISTICREGRESSION(2023-12-07 20:42:57.080333): Seed 51392: run_logistic_regression_multiple_seeds() Running...\n",
      "LOGISTICREGRESSION(2023-12-07 20:47:08.528947): Seed 35667: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 21:00:06.338020, Seed: 51392:\n",
      "{'Accuracy': [0.963900238225156], 'Precision': [0.9065178715833737], 'Recall': [0.8200039012971813], 'F1': [0.8610933299193445], 'FPR': [0.013361896248805597]}\n",
      "LOGISTICREGRESSION(2023-12-07 21:00:06.339713): Seed 51392 done.\n",
      "LOGISTICREGRESSION(2023-12-07 21:00:06.340639): Seed 39263: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 21:01:08.755090, Seed: 35667:\n",
      "{'Accuracy': [0.963900238225156], 'Precision': [0.9066494094806665], 'Recall': [0.8198576026528821], 'F1': [0.8610719864785269], 'FPR': [0.01333877878124711]}\n",
      "LOGISTICREGRESSION(2023-12-07 21:01:08.756667): Seed 35667 done.\n",
      "LOGISTICREGRESSION(2023-12-07 21:01:08.757432): Seed 28115: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 21:19:40.271984, Seed: 28115:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497], 'Precision': [0.9066494094806665, 0.9066630698678176], 'Recall': [0.8198576026528821, 0.8195162391495172], 'F1': [0.8610719864785269, 0.860889833764504], 'FPR': [0.01333877878124711, 0.013331072958727614]}\n",
      "LOGISTICREGRESSION(2023-12-07 21:19:40.291042): Seed 28115 done.\n",
      "LOGISTICREGRESSION(2023-12-07 21:19:40.291754): Seed 39466: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 21:22:54.965848, Seed: 39263:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592], 'Precision': [0.9065178715833737, 0.9066645098673568], 'Recall': [0.8200039012971813, 0.8200039012971813], 'F1': [0.8610933299193445, 0.8611594796681348], 'FPR': [0.013361896248805597, 0.01333877878124711]}\n",
      "LOGISTICREGRESSION(2023-12-07 21:22:54.967550): Seed 39263 done.\n",
      "LOGISTICREGRESSION(2023-12-07 21:22:54.968364): Seed 48764: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 21:41:04.906453, Seed: 48764:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588]}\n",
      "LOGISTICREGRESSION(2023-12-07 21:41:04.908174): Seed 48764 done.\n",
      "LOGISTICREGRESSION(2023-12-07 21:41:04.908903): Seed 30939: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 21:44:22.230734, Seed: 39466:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581]}\n",
      "LOGISTICREGRESSION(2023-12-07 21:44:22.232399): Seed 39466 done.\n",
      "LOGISTICREGRESSION(2023-12-07 21:44:22.233146): Seed 51966: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 22:00:44.335158, Seed: 30939:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497, 0.9638802752232529], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396, 0.90654659188956], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156, 0.8198088364381156], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649, 0.8609987195902689], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588, 0.013354190426286102]}\n",
      "LOGISTICREGRESSION(2023-12-07 22:00:44.336713): Seed 30939 done.\n",
      "LOGISTICREGRESSION(2023-12-07 22:00:44.337322): Seed 10128: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 22:16:57.344109, Seed: 51966:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529, 0.9638669665553175], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207, 0.9065365117031604], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477, 0.8197113040085828], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965, 0.8609403810694529], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581, 0.013354190426286102]}\n",
      "LOGISTICREGRESSION(2023-12-07 22:16:57.345331): Seed 51966 done.\n",
      "LOGISTICREGRESSION(2023-12-07 22:16:57.346495): Seed 57839: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 22:27:26.807283, Seed: 10128:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497, 0.9638802752232529, 0.9638203862175435], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396, 0.90654659188956, 0.9062382056397261], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156, 0.8198088364381156, 0.8196625377938165], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649, 0.8609987195902689, 0.8607789414385579], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588, 0.013354190426286102, 0.013400425361403076]}\n",
      "LOGISTICREGRESSION(2023-12-07 22:27:26.808895): Seed 10128 done.\n",
      "LOGISTICREGRESSION(2023-12-07 22:27:26.809826): Seed 9012: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 22:38:12.986541, Seed: 57839:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529, 0.9638669665553175, 0.9638669665553175], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207, 0.9065365117031604, 0.9066681053085887], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477, 0.8197113040085828, 0.8195650053642837], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965, 0.8609403810694529, 0.8609190102966037], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581, 0.013354190426286102, 0.013331072958727614]}\n",
      "LOGISTICREGRESSION(2023-12-07 22:38:12.988193): Seed 57839 done.\n",
      "LOGISTICREGRESSION(2023-12-07 22:38:12.989103): Seed 6460: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 22:52:08.270092, Seed: 9012:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497, 0.9638802752232529, 0.9638203862175435, 0.9638336948854789], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396, 0.90654659188956, 0.9062382056397261, 0.9062921227152639], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156, 0.8198088364381156, 0.8196625377938165, 0.8197113040085828], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649, 0.8609987195902689, 0.8607789414385579, 0.8608301538934268], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588, 0.013354190426286102, 0.013400425361403076, 0.013392719538883581]}\n",
      "LOGISTICREGRESSION(2023-12-07 22:52:08.271882): Seed 9012 done.\n",
      "LOGISTICREGRESSION(2023-12-07 22:52:08.272617): Seed 3081: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 23:06:14.528539, Seed: 6460:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529, 0.9638669665553175, 0.9638669665553175, 0.963807077549608], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207, 0.9065365117031604, 0.9066681053085887, 0.9061842885641883], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477, 0.8197113040085828, 0.8195650053642837, 0.81961377157905], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965, 0.8609403810694529, 0.8609190102966037, 0.8607277289836889], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581, 0.013354190426286102, 0.013331072958727614, 0.013408131183922572]}\n",
      "LOGISTICREGRESSION(2023-12-07 23:06:14.529716): Seed 6460 done.\n",
      "LOGISTICREGRESSION(2023-12-07 23:06:14.530818): Seed 8562: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 23:10:13.337892, Seed: 3081:\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497, 0.9638802752232529, 0.9638203862175435, 0.9638336948854789, 0.9638736208892852], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396, 0.90654659188956, 0.9062382056397261, 0.9062921227152639, 0.9064100490592485], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156, 0.8198088364381156, 0.8196625377938165, 0.8197113040085828, 0.8199063688676484], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649, 0.8609987195902689, 0.8607789414385579, 0.8608301538934268, 0.8609909102547689], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588, 0.013354190426286102, 0.013400425361403076, 0.013392719538883581, 0.013377307893844588]}\n",
      "LOGISTICREGRESSION(2023-12-07 23:10:13.339566): Seed 3081 done.\n",
      "LOGISTICREGRESSION(2023-12-07 23:10:13.340276): Seed 7647: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 23:26:25.387393, Seed: 8562:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529, 0.9638669665553175, 0.9638669665553175, 0.963807077549608, 0.9638736208892852], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207, 0.9065365117031604, 0.9066681053085887, 0.9061842885641883, 0.9065854053179441], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477, 0.8197113040085828, 0.8195650053642837, 0.81961377157905, 0.8197113040085828], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965, 0.8609403810694529, 0.8609190102966037, 0.8607277289836889, 0.8609624298921812], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581, 0.013354190426286102, 0.013331072958727614, 0.013408131183922572, 0.013346484603766607]}\n",
      "LOGISTICREGRESSION(2023-12-07 23:26:25.388986): Seed 8562 done.\n",
      "LOGISTICREGRESSION(2023-12-07 23:26:25.389821): Seed 18842: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 23:32:37.910442, Seed: 7647:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497, 0.9638802752232529, 0.9638203862175435, 0.9638336948854789, 0.9638736208892852, 0.9638802752232529], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396, 0.90654659188956, 0.9062382056397261, 0.9062921227152639, 0.9064100490592485, 0.9064150943396226], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156, 0.8198088364381156, 0.8196625377938165, 0.8197113040085828, 0.8199063688676484, 0.8199551350824149], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649, 0.8609987195902689, 0.8607789414385579, 0.8608301538934268, 0.8609909102547689, 0.8610200737402703], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588, 0.013354190426286102, 0.013400425361403076, 0.013392719538883581, 0.013377307893844588, 0.013377307893844588]}\n",
      "LOGISTICREGRESSION(2023-12-07 23:32:37.912114): Seed 7647 done.\n",
      "LOGISTICREGRESSION(2023-12-07 23:32:37.912795): Seed 12268: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 23:48:26.863543, Seed: 18842:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529, 0.9638669665553175, 0.9638669665553175, 0.963807077549608, 0.9638736208892852, 0.9638470035534143], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207, 0.9065365117031604, 0.9066681053085887, 0.9061842885641883, 0.9065854053179441, 0.9063898624966298], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477, 0.8197113040085828, 0.8195650053642837, 0.81961377157905, 0.8197113040085828, 0.8197113040085828], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965, 0.8609403810694529, 0.8609190102966037, 0.8607277289836889, 0.8609624298921812, 0.8608742413766612], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581, 0.013354190426286102, 0.013331072958727614, 0.013408131183922572, 0.013346484603766607, 0.013377307893844588]}\n",
      "LOGISTICREGRESSION(2023-12-07 23:48:26.865037): Seed 18842 done.\n",
      "LOGISTICREGRESSION(2023-12-07 23:48:26.865861): Seed 38139: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-07 23:49:01.653315, Seed: 12268:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497, 0.9638802752232529, 0.9638203862175435, 0.9638336948854789, 0.9638736208892852, 0.9638802752232529, 0.9638802752232529], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396, 0.90654659188956, 0.9062382056397261, 0.9062921227152639, 0.9064100490592485, 0.9064150943396226, 0.9066343042071198], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156, 0.8198088364381156, 0.8196625377938165, 0.8197113040085828, 0.8199063688676484, 0.8199551350824149, 0.8197113040085828], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649, 0.8609987195902689, 0.8607789414385579, 0.8608301538934268, 0.8609909102547689, 0.8610200737402703, 0.8609844798442863], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588, 0.013354190426286102, 0.013400425361403076, 0.013392719538883581, 0.013377307893844588, 0.013377307893844588, 0.01333877878124711]}\n",
      "LOGISTICREGRESSION(2023-12-07 23:49:01.655243): Seed 12268 done.\n",
      "LOGISTICREGRESSION(2023-12-07 23:49:01.656092): Seed 31730: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-08 00:05:59.302507, Seed: 38139:\n",
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529, 0.9638669665553175, 0.9638669665553175, 0.963807077549608, 0.9638736208892852, 0.9638470035534143, 0.9638536578873821], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207, 0.9065365117031604, 0.9066681053085887, 0.9061842885641883, 0.9065854053179441, 0.9063898624966298, 0.9064825800884478], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477, 0.8197113040085828, 0.8195650053642837, 0.81961377157905, 0.8197113040085828, 0.8197113040085828, 0.8196625377938165], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965, 0.8609403810694529, 0.8609190102966037, 0.8607277289836889, 0.8609624298921812, 0.8608742413766612, 0.8608891620569556], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581, 0.013354190426286102, 0.013331072958727614, 0.013408131183922572, 0.013346484603766607, 0.013377307893844588, 0.013361896248805597]}\n",
      "LOGISTICREGRESSION(2023-12-08 00:05:59.304138): Seed 38139 done.\n",
      "LOGISTICREGRESSION(2023-12-08 00:05:59.304865): Seed 45917: run_logistic_regression_multiple_seeds() Running...\n",
      "LR55:\n",
      " 2023-12-08 00:12:12.791463, Seed: 31730:\n",
      "{'Accuracy': [0.963900238225156, 0.9639202012270592, 0.9638603122213497, 0.9638802752232529, 0.9638203862175435, 0.9638336948854789, 0.9638736208892852, 0.9638802752232529, 0.9638802752232529, 0.9637871145477049], 'Precision': [0.9065178715833737, 0.9066645098673568, 0.9063999568663396, 0.90654659188956, 0.9062382056397261, 0.9062921227152639, 0.9064100490592485, 0.9064150943396226, 0.9066343042071198, 0.9062129220148851], 'Recall': [0.8200039012971813, 0.8200039012971813, 0.8198088364381156, 0.8198088364381156, 0.8196625377938165, 0.8197113040085828, 0.8199063688676484, 0.8199551350824149, 0.8197113040085828, 0.8194187067199844], 'F1': [0.8610933299193445, 0.8611594796681348, 0.8609325788031649, 0.8609987195902689, 0.8607789414385579, 0.8608301538934268, 0.8609909102547689, 0.8610200737402703, 0.8609844798442863, 0.8606330669944684], 'FPR': [0.013361896248805597, 0.01333877878124711, 0.013377307893844588, 0.013354190426286102, 0.013400425361403076, 0.013392719538883581, 0.013377307893844588, 0.013377307893844588, 0.01333877878124711, 0.013400425361403076]}\n",
      "LOGISTICREGRESSION(2023-12-08 00:12:12.793001): Seed 31730 done.\n",
      "LR55:\n",
      " 2023-12-08 00:21:10.009145, Seed: 45917:\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'Accuracy': [0.963900238225156, 0.9638603122213497, 0.9638802752232529, 0.9638669665553175, 0.9638669665553175, 0.963807077549608, 0.9638736208892852, 0.9638470035534143, 0.9638536578873821, 0.9638470035534143], 'Precision': [0.9066494094806665, 0.9066630698678176, 0.9063274765549207, 0.9065365117031604, 0.9066681053085887, 0.9061842885641883, 0.9065854053179441, 0.9063898624966298, 0.9064825800884478, 0.9063460397908017], 'Recall': [0.8198576026528821, 0.8195162391495172, 0.8200526675119477, 0.8197113040085828, 0.8195650053642837, 0.81961377157905, 0.8197113040085828, 0.8197113040085828, 0.8196625377938165, 0.8197600702233493], 'F1': [0.8610719864785269, 0.860889833764504, 0.8610343061955965, 0.8609403810694529, 0.8609190102966037, 0.8607277289836889, 0.8609624298921812, 0.8608742413766612, 0.8608891620569556, 0.8608813663482959], 'FPR': [0.01333877878124711, 0.013331072958727614, 0.013392719538883581, 0.013354190426286102, 0.013331072958727614, 0.013408131183922572, 0.013346484603766607, 0.013377307893844588, 0.013361896248805597, 0.013385013716364086]}\n",
      "LOGISTICREGRESSION(2023-12-08 00:21:10.011025): Seed 45917 done.\n"
     ]
    }
   ],
   "source": [
    "# Define the tasks\n",
    "\n",
    "base_seed = 42\n",
    "ROUNDS = 10\n",
    "\n",
    "\n",
    "n_workers = 3  # 0 uses all available processors\n",
    "\n",
    "batch1 = {\n",
    "    'Logistic Regression': (run_logistic_regression_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=base_seed)),\n",
    "    'Logistic Regression4_b': (run_logistic_regression_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=142)),\n",
    "    'Logistic Regression4_c': (run_logistic_regression_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=242)),\n",
    "    'Logistic Regression4_d': (run_logistic_regression_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=342)),\n",
    "    'Logistic Regression4_e': (run_logistic_regression_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=442)),\n",
    "    #'SVM': (run_svm_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=base_seed)),\n",
    "    #'DecisionTree': (run_decision_tree_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=base_seed)),\n",
    "    #'RandomForest': (run_random_forest_multiple_seeds, prng_create_seeds(n=ROUNDS,base_seed=base_seed)),\n",
    "}\n",
    "\n",
    "results1 = train_models_in_parallel(batch1, n_workers, X_train, y_train, X_test, y_test)\n",
    "results.update(results1)\n",
    "\n",
    "# XGboost, and LightGBM are multi-threaded by design (running them in threads may cause crashes)\n",
    "#results['AdaBoost'] = run_adaboost_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['XGBoost'] = run_xgb_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n",
    "#results['LightGBM'] = run_lgb_multiple_seeds(prng_create_seeds(n=ROUNDS,base_seed=x), X_train, y_train, X_test, y_test, display_cm=False, display_graphs=False, verbose=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aa2672eb",
   "metadata": {},
   "source": [
    "# Display results and produce outputs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d330d85",
   "metadata": {},
   "source": [
    "## Corret key order"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "ce18fd3d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "b0ab5f6d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def sort_dict_by_key_priority(input_dict, key_priority):\n",
    "    sorted_dict = {key: input_dict[key] for key in key_priority if key in input_dict}\n",
    "    sorted_dict.update({key: input_dict[key] for key in input_dict if key not in key_priority})\n",
    "    return sorted_dict\n",
    "\n",
    "keys = [\"Logistic Regression\", \"SVM\", \"DecisionTree\", \"RandomForest\", \"AdaBoost\", \"XGBoost\", \"LightGBM\"]\n",
    "results = sort_dict_by_key_priority(results, keys)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "771230a9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "febd1afa",
   "metadata": {},
   "source": [
    "## Raw data display"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "6571d35d",
   "metadata": {},
   "outputs": [],
   "source": [
    "results = {\n",
    "'Logistic Regression':\n",
    "{\n",
    "    'Accuracy': [0.9638669665553175,\n",
    "   0.9639202012270592,\n",
    "   0.9638270405515112,\n",
    "   0.9638336948854789,\n",
    "   0.9638536578873821,\n",
    "   0.9638403492194466,\n",
    "   0.9638137318835758,\n",
    "   0.9637804602137372,\n",
    "   0.9638736208892852,\n",
    "   0.963900238225156],\n",
    "  'Precision': [0.9061860114236447,\n",
    "   0.9066645098673568,\n",
    "   0.9061994609164421,\n",
    "   0.906599039551071,\n",
    "   0.9063072776280323,\n",
    "   0.9062971748975631,\n",
    "   0.9064084583018664,\n",
    "   0.9063393579714054,\n",
    "   0.9064100490592485,\n",
    "   0.9067810325295356],\n",
    "  'Recall': [0.8201014337267142,\n",
    "   0.8200039012971813,\n",
    "   0.8197600702233493,\n",
    "   0.8193699405052179,\n",
    "   0.8198576026528821,\n",
    "   0.8197600702233493,\n",
    "   0.8194187067199844,\n",
    "   0.8192236418609188,\n",
    "   0.8199063688676484,\n",
    "   0.8197113040085828],\n",
    "  'F1': [0.8609973377022323,\n",
    "   0.8611594796681348,\n",
    "   0.8608152396558789,\n",
    "   0.86078024539563,\n",
    "   0.8609176566980746,\n",
    "   0.8608593229886823,\n",
    "   0.8607212375781169,\n",
    "   0.8605824645885096,\n",
    "   0.8609909102547689,\n",
    "   0.8610506364777296],\n",
    "  'FPR': [0.013415837006442067,\n",
    "   0.01333877878124711,\n",
    "   0.013408131183922572,\n",
    "   0.01333877878124711,\n",
    "   0.013392719538883581,\n",
    "   0.013392719538883581,\n",
    "   0.013369602071325093,\n",
    "   0.013377307893844588,\n",
    "   0.013377307893844588,\n",
    "   0.013315661313688623]\n",
    "},\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "'SVM': {\n",
    "    \"Accuracy\": [\n",
    "        0.9883482612225343, # base_seed 42 n2\n",
    "        0.9883549155565019,\n",
    "        0.9882950265507925,\n",
    "        0.9882750635488894,\n",
    "        0.9882883722168249,\n",
    "        0.9882484462130185,\n",
    "        0.9882883722168249, # base_seed 142, pchranicky\n",
    "        0.9882817178828571,\n",
    "        0.9882351375450831,\n",
    "        0.9882551005469863  # base_seed 242, pcfitcrack\n",
    "\n",
    "    ],\n",
    "    \"Precision\": [\n",
    "        0.970026565084457,\n",
    "        0.9704524932276513,\n",
    "        0.9697318967677274,\n",
    "        0.9699157641395909,\n",
    "        0.9697303798737096,\n",
    "        0.9692038057085628,\n",
    "        0.9694480617048983,\n",
    "        0.9695876546921188,\n",
    "        0.9693887775551102,\n",
    "        0.9695344991732224\n",
    "\n",
    "    ],\n",
    "    \"Recall\": [\n",
    "        0.9437725543743295,\n",
    "        0.9433824246561981,\n",
    "        0.9436750219447967,\n",
    "        0.9433336584414318,\n",
    "        0.9436262557300302,\n",
    "        0.9438700868038623,\n",
    "        0.9439188530186287,\n",
    "        0.943723788159563,\n",
    "        0.9435774895152639,\n",
    "        0.9435774895152639\n",
    "\n",
    "    ],\n",
    "    \"F1\": [\n",
    "        0.9567194799416665,\n",
    "        0.9567260138476755,\n",
    "        0.9565260374187491,\n",
    "        0.9564400494437578,\n",
    "        0.9565002471576867,\n",
    "        0.9563692064433243,\n",
    "        0.9565131448902945,\n",
    "        0.9564809094279006,\n",
    "        0.9563090001482726,\n",
    "        0.9568226193124244\n",
    "\n",
    "    ],\n",
    "    \"FPR\": [\n",
    "        0.004608081866658447,\n",
    "        0.004538729463982986,\n",
    "        0.004654316801775422,\n",
    "        0.004623493511697439,\n",
    "        0.004654316801775422,\n",
    "        0.004739080849489875,\n",
    "        0.004700551736892396,\n",
    "        0.004677434269333909,\n",
    "        0.004708257559411891,\n",
    "        0.004685140091853404\n",
    "    ]\n",
    "},\n",
    "\n",
    "\n",
    "'DecisionTree': {\n",
    "  'Accuracy': [0.9824991016649144,\n",
    "   0.9824991016649144,\n",
    "   0.9825057559988821,\n",
    "   0.9824991016649144,\n",
    "   0.9825589906706238,\n",
    "   0.9824991016649144,\n",
    "   0.9825057559988821,\n",
    "   0.9825257190007852,\n",
    "   0.9824857929969789,\n",
    "   0.9825057559988821],\n",
    "  'Precision': [0.9650364203954215,\n",
    "   0.9651332223147377,\n",
    "   0.965183450429352,\n",
    "   0.9650848163180352,\n",
    "   0.9657794676806084,\n",
    "   0.9649880345437519,\n",
    "   0.9654256703983337,\n",
    "   0.96538261322228,\n",
    "   0.9651780137414117,\n",
    "   0.9650866330194079],\n",
    "  'Recall': [0.9045157514873695,\n",
    "   0.9044182190578367,\n",
    "   0.9044182190578367,\n",
    "   0.9044669852726032,\n",
    "   0.9042231541987711,\n",
    "   0.904564517702136,\n",
    "   0.9041743879840047,\n",
    "   0.9043694528430704,\n",
    "   0.9042719204135375,\n",
    "   0.9045157514873695],\n",
    "  'F1': [0.933796506066556,\n",
    "   0.9337898393837168,\n",
    "   0.9338133481030186,\n",
    "   0.9337931728929615,\n",
    "   0.9339881626999119,\n",
    "   0.933799838904551,\n",
    "   0.9337966810203722,\n",
    "   0.9338805519186223,\n",
    "   0.9337328163553049,\n",
    "   0.9338200125865324],\n",
    "  'FPR': [0.005178312733101131,\n",
    "   0.00516290108806214,\n",
    "   0.005155195265542644,\n",
    "   0.005170606910581636,\n",
    "   0.005062725395308695,\n",
    "   0.005186018555620627,\n",
    "   0.005116666152945165,\n",
    "   0.005124371975464661,\n",
    "   0.005155195265542644,\n",
    "   0.005170606910581636]\n",
    "},\n",
    "\n",
    "'RandomForest': {\n",
    "    \"Accuracy\": [\n",
    "        0.9846950318742597,\n",
    "        0.9846284885345826,\n",
    "        0.9844687845193575,\n",
    "        0.9846684145383888,\n",
    "        0.9847016862082274,\n",
    "        0.9845619451949055,\n",
    "        0.9844621301853897,\n",
    "        0.9846484515364857,\n",
    "        0.9846284885345826,\n",
    "        0.9845819081968086\n",
    "    ],\n",
    "    \"Precision\": [\n",
    "        0.9772965604026845,\n",
    "        0.9779867605337816,\n",
    "        0.9775068320369981,\n",
    "        0.977742782152231,\n",
    "        0.9783008458992276,\n",
    "        0.9773729525409491,\n",
    "        0.977355120054642,\n",
    "        0.9779902295529758,\n",
    "        0.9774349286314021,\n",
    "        0.977677399023058\n",
    "    ],\n",
    "    \"Recall\": [\n",
    "        0.9089534770311128,\n",
    "        0.907783087876719,\n",
    "        0.9070515946552229,\n",
    "        0.9083195162391495,\n",
    "        0.9080269189505511,\n",
    "        0.9078806203062518,\n",
    "        0.9071491270847557,\n",
    "        0.9079293865210183,\n",
    "        0.9083195162391495,\n",
    "        0.9077343216619526\n",
    "    ],\n",
    "    \"F1\": [\n",
    "        0.9418869068674516,\n",
    "        0.9415781487101669,\n",
    "        0.940962209743512,\n",
    "        0.941753463444231,\n",
    "        0.9418548774627583,\n",
    "        0.9413460079890783,\n",
    "        0.9409443840258985,\n",
    "        0.9416584477657235,\n",
    "        0.9416106364693392,\n",
    "        0.9414085219370337\n",
    "    ],\n",
    "    \"FPR\": [\n",
    "        0.0033366211509416517,\n",
    "        0.0032287396356687112,\n",
    "        0.0032980920383441728,\n",
    "        0.0032672687482661897,\n",
    "        0.003182504700551737,\n",
    "        0.00332120950590266,\n",
    "        0.00332120950590266,\n",
    "        0.0032287396356687112,\n",
    "        0.0033135036833831645,\n",
    "        0.0032749745707856856\n",
    "    ]\n",
    "},\n",
    "\n",
    "'AdaBoost':\n",
    "   {'Accuracy': [0.9902181290674616,\n",
    "   0.9902247834014294,\n",
    "   0.9902447464033325,\n",
    "   0.9902447464033325,\n",
    "   0.990231437735397,\n",
    "   0.9902447464033325,\n",
    "   0.9902447464033325,\n",
    "   0.9902181290674616,\n",
    "   0.9902380920693648,\n",
    "   0.990231437735397],\n",
    "  'Precision': [0.9705824186690398,\n",
    "   0.9705838730409848,\n",
    "   0.9707278481012658,\n",
    "   0.9707278481012658,\n",
    "   0.9707249530214618,\n",
    "   0.9707278481012658,\n",
    "   0.9707278481012658,\n",
    "   0.9705358908443741,\n",
    "   0.9706798516687268,\n",
    "   0.9707249530214618],\n",
    "  'Recall': [0.9573295620793913,\n",
    "   0.9573783282941578,\n",
    "   0.9573783282941578,\n",
    "   0.9573783282941578,\n",
    "   0.957280795864625,\n",
    "   0.9573783282941578,\n",
    "   0.9573783282941578,\n",
    "   0.9573783282941578,\n",
    "   0.9573783282941578,\n",
    "   0.957280795864625],\n",
    "  'F1': [0.9639104389669055,\n",
    "   0.963935875088994,\n",
    "   0.9640068745396513,\n",
    "   0.9640068745396513,\n",
    "   0.9639560007857003,\n",
    "   0.9640068745396513,\n",
    "   0.9640068745396513,\n",
    "   0.9639122109294447,\n",
    "   0.963983206894012,\n",
    "   0.9639560007857003],\n",
    "  'FPR': [0.00458496439909996,\n",
    "   0.00458496439909996,\n",
    "   0.004561846931541473,\n",
    "   0.004561846931541473,\n",
    "   0.004561846931541473,\n",
    "   0.004561846931541473,\n",
    "   0.004561846931541473,\n",
    "   0.0045926702216194555,\n",
    "   0.004569552754060968,\n",
    "   0.004561846931541473]},\n",
    "\n",
    "'XGBoost': {\n",
    "  'Accuracy': [0.9936251480589308,\n",
    "   0.9936850370646402,\n",
    "   0.9934920613795765,\n",
    "   0.9936517653948016,\n",
    "   0.993645111060834,\n",
    "   0.993551950385286,\n",
    "   0.9936850370646402,\n",
    "   0.9936783827306724,\n",
    "   0.9936318023928985,\n",
    "   0.9935785677211568],\n",
    "  'Precision': [0.9877245508982035,\n",
    "   0.9878273883761537,\n",
    "   0.9871283177010577,\n",
    "   0.9878243512974052,\n",
    "   0.9877750611246944,\n",
    "   0.9876204262966106,\n",
    "   0.9878273883761537,\n",
    "   0.9875349022736338,\n",
    "   0.9879199321120151,\n",
    "   0.987476924612084],\n",
    "  'Recall': [0.9652784550863162,\n",
    "   0.965619818589681,\n",
    "   0.9648883253681849,\n",
    "   0.965375987515849,\n",
    "   0.965375987515849,\n",
    "   0.9648395591534185,\n",
    "   0.965619818589681,\n",
    "   0.9658636496635131,\n",
    "   0.965132156442017,\n",
    "   0.9651809226567833],\n",
    "  'F1': [0.9763725151679573,\n",
    "   0.9765973712115608,\n",
    "   0.9758816276202219,\n",
    "   0.9764711685493019,\n",
    "   0.9764470860976151,\n",
    "   0.9760970916894842,\n",
    "   0.9765973712115608,\n",
    "   0.9765790641487105,\n",
    "   0.9763931029379117,\n",
    "   0.9762016325926656],\n",
    "  'FPR': [0.0018956323397959498,\n",
    "   0.0018802206947569583,\n",
    "   0.0019881022100298987,\n",
    "   0.0018802206947569583,\n",
    "   0.0018879265172764541,\n",
    "   0.0019110439848349413,\n",
    "   0.0018802206947569583,\n",
    "   0.0019264556298739328,\n",
    "   0.001864809049717967,\n",
    "   0.0019341614523934285]},\n",
    "\n",
    "'LightGBM': {\n",
    "'Accuracy': [0.993944556089381,\n",
    "   0.9940044450950904,\n",
    "   0.9940976057706384,\n",
    "   0.993991136427155,\n",
    "   0.994037716764929,\n",
    "   0.9939778277592196,\n",
    "   0.9939778277592196,\n",
    "   0.993851395413833,\n",
    "   0.9938181237439945,\n",
    "   0.9940443710988968],\n",
    "  'Precision': [0.9881426863292149,\n",
    "   0.988829601555877,\n",
    "   0.9891786765072558,\n",
    "   0.9885848163102537,\n",
    "   0.9886374962623343,\n",
    "   0.9885836781494591,\n",
    "   0.9891689543299227,\n",
    "   0.9884266187768134,\n",
    "   0.9883262659017211,\n",
    "   0.9883460331689825],\n",
    "  'Recall': [0.9672291036769726,\n",
    "   0.9669852726031405,\n",
    "   0.9673266361065054,\n",
    "   0.9671315712474398,\n",
    "   0.9674241685360382,\n",
    "   0.967034038817907,\n",
    "   0.96644884424071,\n",
    "   0.9662537793816444,\n",
    "   0.9661074807373452,\n",
    "   0.9677655320394031],\n",
    "  'F1': [0.9775740549065997,\n",
    "   0.9777854483591806,\n",
    "   0.9781306245223008,\n",
    "   0.977740528015382,\n",
    "   0.9779158040027606,\n",
    "   0.9776901269567361,\n",
    "   0.9776769196615771,\n",
    "   0.9772144407180904,\n",
    "   0.9770905772977239,\n",
    "   0.9779475175557472],\n",
    "  'FPR': [0.001833985759639984,\n",
    "   0.0017261042443670438,\n",
    "   0.0016721634867305735,\n",
    "   0.0017646333569645224,\n",
    "   0.0017569275344450266,\n",
    "   0.0017646333569645224,\n",
    "   0.0016721634867305735,\n",
    "   0.0017877508245230096,\n",
    "   0.0018031624695620011,\n",
    "   0.0018031624695620011]},\n",
    "\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "bef09d24",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Logistic Regression': {'Accuracy': [0.9638669665553175,\n",
       "   0.9639202012270592,\n",
       "   0.9638270405515112,\n",
       "   0.9638336948854789,\n",
       "   0.9638536578873821,\n",
       "   0.9638403492194466,\n",
       "   0.9638137318835758,\n",
       "   0.9637804602137372,\n",
       "   0.9638736208892852,\n",
       "   0.963900238225156],\n",
       "  'Precision': [0.9061860114236447,\n",
       "   0.9066645098673568,\n",
       "   0.9061994609164421,\n",
       "   0.906599039551071,\n",
       "   0.9063072776280323,\n",
       "   0.9062971748975631,\n",
       "   0.9064084583018664,\n",
       "   0.9063393579714054,\n",
       "   0.9064100490592485,\n",
       "   0.9067810325295356],\n",
       "  'Recall': [0.8201014337267142,\n",
       "   0.8200039012971813,\n",
       "   0.8197600702233493,\n",
       "   0.8193699405052179,\n",
       "   0.8198576026528821,\n",
       "   0.8197600702233493,\n",
       "   0.8194187067199844,\n",
       "   0.8192236418609188,\n",
       "   0.8199063688676484,\n",
       "   0.8197113040085828],\n",
       "  'F1': [0.8609973377022323,\n",
       "   0.8611594796681348,\n",
       "   0.8608152396558789,\n",
       "   0.86078024539563,\n",
       "   0.8609176566980746,\n",
       "   0.8608593229886823,\n",
       "   0.8607212375781169,\n",
       "   0.8605824645885096,\n",
       "   0.8609909102547689,\n",
       "   0.8610506364777296],\n",
       "  'FPR': [0.013415837006442067,\n",
       "   0.01333877878124711,\n",
       "   0.013408131183922572,\n",
       "   0.01333877878124711,\n",
       "   0.013392719538883581,\n",
       "   0.013392719538883581,\n",
       "   0.013369602071325093,\n",
       "   0.013377307893844588,\n",
       "   0.013377307893844588,\n",
       "   0.013315661313688623]},\n",
       " 'SVM': {'Accuracy': [0.9883482612225343,\n",
       "   0.9883549155565019,\n",
       "   0.9882950265507925,\n",
       "   0.9882750635488894,\n",
       "   0.9882883722168249,\n",
       "   0.9882484462130185,\n",
       "   0.9882883722168249,\n",
       "   0.9882817178828571,\n",
       "   0.9882351375450831,\n",
       "   0.9882551005469863],\n",
       "  'Precision': [0.970026565084457,\n",
       "   0.9704524932276513,\n",
       "   0.9697318967677274,\n",
       "   0.9699157641395909,\n",
       "   0.9697303798737096,\n",
       "   0.9692038057085628,\n",
       "   0.9694480617048983,\n",
       "   0.9695876546921188,\n",
       "   0.9693887775551102,\n",
       "   0.9695344991732224],\n",
       "  'Recall': [0.9437725543743295,\n",
       "   0.9433824246561981,\n",
       "   0.9436750219447967,\n",
       "   0.9433336584414318,\n",
       "   0.9436262557300302,\n",
       "   0.9438700868038623,\n",
       "   0.9439188530186287,\n",
       "   0.943723788159563,\n",
       "   0.9435774895152639,\n",
       "   0.9435774895152639],\n",
       "  'F1': [0.9567194799416665,\n",
       "   0.9567260138476755,\n",
       "   0.9565260374187491,\n",
       "   0.9564400494437578,\n",
       "   0.9565002471576867,\n",
       "   0.9563692064433243,\n",
       "   0.9565131448902945,\n",
       "   0.9564809094279006,\n",
       "   0.9563090001482726,\n",
       "   0.9568226193124244],\n",
       "  'FPR': [0.004608081866658447,\n",
       "   0.004538729463982986,\n",
       "   0.004654316801775422,\n",
       "   0.004623493511697439,\n",
       "   0.004654316801775422,\n",
       "   0.004739080849489875,\n",
       "   0.004700551736892396,\n",
       "   0.004677434269333909,\n",
       "   0.004708257559411891,\n",
       "   0.004685140091853404]},\n",
       " 'DecisionTree': {'Accuracy': [0.9824991016649144,\n",
       "   0.9824991016649144,\n",
       "   0.9825057559988821,\n",
       "   0.9824991016649144,\n",
       "   0.9825589906706238,\n",
       "   0.9824991016649144,\n",
       "   0.9825057559988821,\n",
       "   0.9825257190007852,\n",
       "   0.9824857929969789,\n",
       "   0.9825057559988821],\n",
       "  'Precision': [0.9650364203954215,\n",
       "   0.9651332223147377,\n",
       "   0.965183450429352,\n",
       "   0.9650848163180352,\n",
       "   0.9657794676806084,\n",
       "   0.9649880345437519,\n",
       "   0.9654256703983337,\n",
       "   0.96538261322228,\n",
       "   0.9651780137414117,\n",
       "   0.9650866330194079],\n",
       "  'Recall': [0.9045157514873695,\n",
       "   0.9044182190578367,\n",
       "   0.9044182190578367,\n",
       "   0.9044669852726032,\n",
       "   0.9042231541987711,\n",
       "   0.904564517702136,\n",
       "   0.9041743879840047,\n",
       "   0.9043694528430704,\n",
       "   0.9042719204135375,\n",
       "   0.9045157514873695],\n",
       "  'F1': [0.933796506066556,\n",
       "   0.9337898393837168,\n",
       "   0.9338133481030186,\n",
       "   0.9337931728929615,\n",
       "   0.9339881626999119,\n",
       "   0.933799838904551,\n",
       "   0.9337966810203722,\n",
       "   0.9338805519186223,\n",
       "   0.9337328163553049,\n",
       "   0.9338200125865324],\n",
       "  'FPR': [0.005178312733101131,\n",
       "   0.00516290108806214,\n",
       "   0.005155195265542644,\n",
       "   0.005170606910581636,\n",
       "   0.005062725395308695,\n",
       "   0.005186018555620627,\n",
       "   0.005116666152945165,\n",
       "   0.005124371975464661,\n",
       "   0.005155195265542644,\n",
       "   0.005170606910581636]},\n",
       " 'RandomForest': {'Accuracy': [0.9846950318742597,\n",
       "   0.9846284885345826,\n",
       "   0.9844687845193575,\n",
       "   0.9846684145383888,\n",
       "   0.9847016862082274,\n",
       "   0.9845619451949055,\n",
       "   0.9844621301853897,\n",
       "   0.9846484515364857,\n",
       "   0.9846284885345826,\n",
       "   0.9845819081968086],\n",
       "  'Precision': [0.9772965604026845,\n",
       "   0.9779867605337816,\n",
       "   0.9775068320369981,\n",
       "   0.977742782152231,\n",
       "   0.9783008458992276,\n",
       "   0.9773729525409491,\n",
       "   0.977355120054642,\n",
       "   0.9779902295529758,\n",
       "   0.9774349286314021,\n",
       "   0.977677399023058],\n",
       "  'Recall': [0.9089534770311128,\n",
       "   0.907783087876719,\n",
       "   0.9070515946552229,\n",
       "   0.9083195162391495,\n",
       "   0.9080269189505511,\n",
       "   0.9078806203062518,\n",
       "   0.9071491270847557,\n",
       "   0.9079293865210183,\n",
       "   0.9083195162391495,\n",
       "   0.9077343216619526],\n",
       "  'F1': [0.9418869068674516,\n",
       "   0.9415781487101669,\n",
       "   0.940962209743512,\n",
       "   0.941753463444231,\n",
       "   0.9418548774627583,\n",
       "   0.9413460079890783,\n",
       "   0.9409443840258985,\n",
       "   0.9416584477657235,\n",
       "   0.9416106364693392,\n",
       "   0.9414085219370337],\n",
       "  'FPR': [0.0033366211509416517,\n",
       "   0.0032287396356687112,\n",
       "   0.0032980920383441728,\n",
       "   0.0032672687482661897,\n",
       "   0.003182504700551737,\n",
       "   0.00332120950590266,\n",
       "   0.00332120950590266,\n",
       "   0.0032287396356687112,\n",
       "   0.0033135036833831645,\n",
       "   0.0032749745707856856]},\n",
       " 'AdaBoost': {'Accuracy': [0.9902181290674616,\n",
       "   0.9902247834014294,\n",
       "   0.9902447464033325,\n",
       "   0.9902447464033325,\n",
       "   0.990231437735397,\n",
       "   0.9902447464033325,\n",
       "   0.9902447464033325,\n",
       "   0.9902181290674616,\n",
       "   0.9902380920693648,\n",
       "   0.990231437735397],\n",
       "  'Precision': [0.9705824186690398,\n",
       "   0.9705838730409848,\n",
       "   0.9707278481012658,\n",
       "   0.9707278481012658,\n",
       "   0.9707249530214618,\n",
       "   0.9707278481012658,\n",
       "   0.9707278481012658,\n",
       "   0.9705358908443741,\n",
       "   0.9706798516687268,\n",
       "   0.9707249530214618],\n",
       "  'Recall': [0.9573295620793913,\n",
       "   0.9573783282941578,\n",
       "   0.9573783282941578,\n",
       "   0.9573783282941578,\n",
       "   0.957280795864625,\n",
       "   0.9573783282941578,\n",
       "   0.9573783282941578,\n",
       "   0.9573783282941578,\n",
       "   0.9573783282941578,\n",
       "   0.957280795864625],\n",
       "  'F1': [0.9639104389669055,\n",
       "   0.963935875088994,\n",
       "   0.9640068745396513,\n",
       "   0.9640068745396513,\n",
       "   0.9639560007857003,\n",
       "   0.9640068745396513,\n",
       "   0.9640068745396513,\n",
       "   0.9639122109294447,\n",
       "   0.963983206894012,\n",
       "   0.9639560007857003],\n",
       "  'FPR': [0.00458496439909996,\n",
       "   0.00458496439909996,\n",
       "   0.004561846931541473,\n",
       "   0.004561846931541473,\n",
       "   0.004561846931541473,\n",
       "   0.004561846931541473,\n",
       "   0.004561846931541473,\n",
       "   0.0045926702216194555,\n",
       "   0.004569552754060968,\n",
       "   0.004561846931541473]},\n",
       " 'XGBoost': {'Accuracy': [0.9936251480589308,\n",
       "   0.9936850370646402,\n",
       "   0.9934920613795765,\n",
       "   0.9936517653948016,\n",
       "   0.993645111060834,\n",
       "   0.993551950385286,\n",
       "   0.9936850370646402,\n",
       "   0.9936783827306724,\n",
       "   0.9936318023928985,\n",
       "   0.9935785677211568],\n",
       "  'Precision': [0.9877245508982035,\n",
       "   0.9878273883761537,\n",
       "   0.9871283177010577,\n",
       "   0.9878243512974052,\n",
       "   0.9877750611246944,\n",
       "   0.9876204262966106,\n",
       "   0.9878273883761537,\n",
       "   0.9875349022736338,\n",
       "   0.9879199321120151,\n",
       "   0.987476924612084],\n",
       "  'Recall': [0.9652784550863162,\n",
       "   0.965619818589681,\n",
       "   0.9648883253681849,\n",
       "   0.965375987515849,\n",
       "   0.965375987515849,\n",
       "   0.9648395591534185,\n",
       "   0.965619818589681,\n",
       "   0.9658636496635131,\n",
       "   0.965132156442017,\n",
       "   0.9651809226567833],\n",
       "  'F1': [0.9763725151679573,\n",
       "   0.9765973712115608,\n",
       "   0.9758816276202219,\n",
       "   0.9764711685493019,\n",
       "   0.9764470860976151,\n",
       "   0.9760970916894842,\n",
       "   0.9765973712115608,\n",
       "   0.9765790641487105,\n",
       "   0.9763931029379117,\n",
       "   0.9762016325926656],\n",
       "  'FPR': [0.0018956323397959498,\n",
       "   0.0018802206947569583,\n",
       "   0.0019881022100298987,\n",
       "   0.0018802206947569583,\n",
       "   0.0018879265172764541,\n",
       "   0.0019110439848349413,\n",
       "   0.0018802206947569583,\n",
       "   0.0019264556298739328,\n",
       "   0.001864809049717967,\n",
       "   0.0019341614523934285]},\n",
       " 'LightGBM': {'Accuracy': [0.993944556089381,\n",
       "   0.9940044450950904,\n",
       "   0.9940976057706384,\n",
       "   0.993991136427155,\n",
       "   0.994037716764929,\n",
       "   0.9939778277592196,\n",
       "   0.9939778277592196,\n",
       "   0.993851395413833,\n",
       "   0.9938181237439945,\n",
       "   0.9940443710988968],\n",
       "  'Precision': [0.9881426863292149,\n",
       "   0.988829601555877,\n",
       "   0.9891786765072558,\n",
       "   0.9885848163102537,\n",
       "   0.9886374962623343,\n",
       "   0.9885836781494591,\n",
       "   0.9891689543299227,\n",
       "   0.9884266187768134,\n",
       "   0.9883262659017211,\n",
       "   0.9883460331689825],\n",
       "  'Recall': [0.9672291036769726,\n",
       "   0.9669852726031405,\n",
       "   0.9673266361065054,\n",
       "   0.9671315712474398,\n",
       "   0.9674241685360382,\n",
       "   0.967034038817907,\n",
       "   0.96644884424071,\n",
       "   0.9662537793816444,\n",
       "   0.9661074807373452,\n",
       "   0.9677655320394031],\n",
       "  'F1': [0.9775740549065997,\n",
       "   0.9777854483591806,\n",
       "   0.9781306245223008,\n",
       "   0.977740528015382,\n",
       "   0.9779158040027606,\n",
       "   0.9776901269567361,\n",
       "   0.9776769196615771,\n",
       "   0.9772144407180904,\n",
       "   0.9770905772977239,\n",
       "   0.9779475175557472],\n",
       "  'FPR': [0.001833985759639984,\n",
       "   0.0017261042443670438,\n",
       "   0.0016721634867305735,\n",
       "   0.0017646333569645224,\n",
       "   0.0017569275344450266,\n",
       "   0.0017646333569645224,\n",
       "   0.0016721634867305735,\n",
       "   0.0017877508245230096,\n",
       "   0.0018031624695620011,\n",
       "   0.0018031624695620011]}}"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f8c4fa67",
   "metadata": {},
   "source": [
    "## Draw table with averages and variances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "323380db",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'table_with_variances.png'"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np2\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def calculate_averages_and_variances(results):\n",
    "    stats = {}\n",
    "    for classifier, metrics in results.items():\n",
    "        stats[classifier] = {}\n",
    "        for metric, values in metrics.items():\n",
    "            avg = np.mean(values)\n",
    "            var = np.var(values, ddof=1)  # Unbiased estimator, using sample variance\n",
    "            stats[classifier][metric] = {'average': avg, 'variance': var}\n",
    "    return stats\n",
    "\n",
    "def draw_table_with_variances(stats):\n",
    "    # Define the metrics\n",
    "    metrics = ['Precision', 'Recall', 'F1', 'FPR']\n",
    "    # Create headers for the table: 'Metric (Avg.)' followed by 'Metric (Variance)'\n",
    "    headers = ['Classifier']\n",
    "    for metric in metrics:\n",
    "        headers.extend([f'{metric} (Avg.)', f'{metric} (Variance)'])\n",
    "    \n",
    "    # Create the table data structure\n",
    "    cell_text = [headers]\n",
    "    \n",
    "    # Fill in the data rows\n",
    "    for classifier, metrics_data in stats.items():\n",
    "        row = [classifier]\n",
    "        for metric in metrics:\n",
    "            avg = metrics_data[metric]['average']\n",
    "            var = metrics_data[metric]['variance']\n",
    "            row.extend([f\"{avg:.6f}\", f\"{var:.2e}\"])\n",
    "        cell_text.append(row)\n",
    "\n",
    "    # Determine the figure width dynamically: base_width + (num_columns * width_per_column)\n",
    "    num_columns = len(headers)\n",
    "    width_per_column = 1.0  # starting point for width per column, may need adjustment\n",
    "    base_width = 4  # base width to account for the first column which typically needs more space\n",
    "    fig_width = base_width + (num_columns * width_per_column)\n",
    "\n",
    "    # Create the figure and axis\n",
    "    fig, ax = plt.subplots(figsize=(fig_width, 2 + 0.625 * len(cell_text)))  # Adjust the size as needed\n",
    "    ax.axis('off')  # Hide the axes\n",
    "\n",
    "    # Create the table and adjust the layout\n",
    "    table = ax.table(cellText=cell_text, loc='center', cellLoc='center')\n",
    "    table.auto_set_font_size(False)\n",
    "    table.set_fontsize(10)\n",
    "\n",
    "    plt.tight_layout()\n",
    "    return fig\n",
    "\n",
    "# Calculate averages and variances\n",
    "stats = calculate_averages_and_variances(results)\n",
    "\n",
    "# Draw the table with variances\n",
    "fig = draw_table_with_variances(stats)\n",
    "\n",
    "# Save the figure to a file and show it\n",
    "output_file = \"table_with_variances.png\"\n",
    "fig.savefig(output_file)\n",
    "plt.show()\n",
    "\n",
    "# Return the path to the saved figure\n",
    "output_file\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4f8138b",
   "metadata": {},
   "source": [
    "## Shorthands for classifier names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "f01bf503",
   "metadata": {},
   "outputs": [],
   "source": [
    "shorthands = {\n",
    "    'Logistic Regression': 'LR',\n",
    "    'DecisionTree': 'DT',\n",
    "    'RandomForest': 'RF',\n",
    "    'SVM': 'SVM',\n",
    "    'AdaBoost': 'ADAB',\n",
    "    'XGBoost': 'XGB',\n",
    "    'LightGBM': 'LGBM',\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9529ab6a",
   "metadata": {},
   "source": [
    "## Generate LaTeX tabular code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "a41696e2",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{|l|c|c|c|c|c|c|c|c|}\n",
      "\\hline\n",
      " & \\multicolumn{2}{c|}{\\textbf{Precision}} & \\multicolumn{2}{c|}{\\textbf{Recall}} & \\multicolumn{2}{c|}{\\textbf{F1}} & \\multicolumn{2}{c|}{\\textbf{FPR}} \\\\\n",
      "\\hline\n",
      "\\textbf{Classifier} & Avg. & Variance & Avg. & Variance & Avg. & Variance & Avg. & Variance \\\\\n",
      "\\hline\n",
      "Logistic Regression (LR) & 0.906419 & 4.00e-08 & 0.819711 & 8.24e-08 & 0.860887 & 2.92e-08 & 0.013373 & 1.06e-09 \\\\\n",
      "\\hline\n",
      "SVM & 0.969702 & 1.30e-07 & 0.943646 & 3.60e-08 & 0.956541 & 2.72e-08 & 0.004659 & 3.33e-09 \\\\\n",
      "\\hline\n",
      "DecisionTree (DT) & 0.965228 & 5.73e-08 & 0.904394 & 1.76e-08 & 0.933821 & 4.75e-09 & 0.005148 & 1.39e-09 \\\\\n",
      "\\hline\n",
      "RandomForest (RF) & 0.977666 & 1.13e-07 & 0.907915 & 3.11e-07 & 0.941500 & 1.13e-07 & 0.003277 & 2.55e-09 \\\\\n",
      "\\hline\n",
      "AdaBoost (ADAB) & 0.970674 & 5.82e-09 & 0.957354 & 1.72e-09 & 0.963968 & 1.56e-09 & 0.004570 & 1.51e-10 \\\\\n",
      "\\hline\n",
      "XGBoost (XGB) & 0.987666 & 5.57e-08 & 0.965317 & 1.06e-07 & 0.976364 & 5.61e-08 & 0.001905 & 1.34e-09 \\\\\n",
      "\\hline\n",
      "LightGBM (LGBM) & \\textbf{0.988622} & 1.21e-07 & \\textbf{0.966971} & 2.88e-07 & \\textbf{0.977677} & 1.02e-07 & \\textbf{0.001758} & 2.95e-09 \\\\\n",
      "\\hline\n",
      "\\end{tabular}\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'table_with_variances.tex'"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def generate_latex_tabular(stats):\n",
    "    # Define the metrics\n",
    "    metrics = ['Precision', 'Recall', 'F1', 'FPR']\n",
    "    \n",
    "    # Find the highest average for each metric except for FPR, where we find the lowest\n",
    "    highest_avg = {metric: max(metrics_data[metric]['average'] for metrics_data in stats.values()) for metric in metrics if metric != 'FPR'}\n",
    "    lowest_fpr = min(stats[classifier]['FPR']['average'] for classifier in stats)\n",
    "    \n",
    "    # Initialize the LaTeX table string\n",
    "    latex_table = \"\\\\begin{tabular}{|l|\" + \"c|c|\" * len(metrics) + \"}\\n\\\\hline\\n\"\n",
    "    \n",
    "    # Create the header row with multicolumns for average and variance\n",
    "    latex_table += \" & \" + \" & \".join([f\"\\\\multicolumn{{2}}{{c|}}{{\\\\textbf{{{metric}}}}}\" for metric in metrics]) + \" \\\\\\\\\\n\"\n",
    "    latex_table += \"\\\\hline\\n\"\n",
    "    latex_table += \"\\\\textbf{Classifier} & \" + \" & \".join([\"Avg. & Variance\" for _ in metrics]) + \" \\\\\\\\\\n\\\\hline\\n\"\n",
    "    \n",
    "    # Fill in the data rows\n",
    "    for classifier, metrics_data in stats.items():\n",
    "        cln = classifier\n",
    "        shorthand = shorthands[classifier]\n",
    "        if shorthand != classifier:\n",
    "            cln = f\"{classifier} ({shorthand})\"\n",
    "        row = [cln.replace('_', '\\\\_')]  # Escape underscores for LaTeX\n",
    "        for metric in metrics:\n",
    "            avg = metrics_data[metric]['average']\n",
    "            # Bold the highest average or the lowest FPR\n",
    "            if metric == 'FPR':\n",
    "                avg_str = f\"\\\\textbf{{{avg:.6f}}}\" if avg == lowest_fpr else f\"{avg:.6f}\"\n",
    "            else:\n",
    "                avg_str = f\"\\\\textbf{{{avg:.6f}}}\" if avg == highest_avg.get(metric, None) else f\"{avg:.6f}\"\n",
    "            var = metrics_data[metric]['variance']\n",
    "            # Display variance in scientific notation\n",
    "            var_str = \"{:.2e}\".format(var) if var > 0 else \"0.00e+00\"\n",
    "            row.extend([avg_str, var_str])\n",
    "        latex_table += \" & \".join(row) + \" \\\\\\\\\\n\\\\hline\\n\"\n",
    "    \n",
    "    # Close the LaTeX table string\n",
    "    latex_table += \"\\\\end{tabular}\"\n",
    "    \n",
    "    return latex_table\n",
    "\n",
    "# Calculate averages and variances\n",
    "stats = calculate_averages_and_variances(results)\n",
    "\n",
    "# Generate LaTeX tabular code for the table\n",
    "latex_tabular_code = generate_latex_tabular(stats)\n",
    "print(latex_tabular_code)\n",
    "\n",
    "# Save the LaTeX code to a file\n",
    "output_file = \"table_with_variances.tex\"\n",
    "with open(output_file, \"w\") as f:\n",
    "    f.write(latex_tabular_code)\n",
    "\n",
    "# Return the path to the saved LaTeX file\n",
    "output_file\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ad9e6b5d",
   "metadata": {},
   "source": [
    "## Boxplots for boosted trees only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "36a09ce6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x700 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import copy\n",
    "\n",
    "#boosted_trees = [\"AdaBoost\", \"XGBoost\", \"LightGBM\"]\n",
    "boosted_trees = [\"XGBoost\", \"LightGBM\"]\n",
    "results_filtered = {shorthands[classifier]: results[classifier] for classifier in boosted_trees if classifier in results}\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Function to adjust the y-axis limits\n",
    "def get_ylim(data_list):\n",
    "    data_min = np.min(data_list)\n",
    "    data_max = np.max(data_list)\n",
    "    \n",
    "    # Inflate the range slightly to make sure all data points are comfortably within the plot\n",
    "    padding = (data_max - data_min) * 0.05\n",
    "    lower = data_min - padding\n",
    "    upper = data_max + padding\n",
    "    \n",
    "    # Ensure that the lower limit is not less than 0 since these metrics cannot be negative\n",
    "    lower = max(lower, 0)\n",
    "    return (lower, upper)\n",
    "\n",
    "\n",
    "# Adjust the plotting function to display results for all classifiers\n",
    "def plot_metric(metric, results, ax):\n",
    "    data = [results[classifier][metric] for classifier in results]\n",
    "    ax.boxplot(data, positions=range(1, len(results)+1), widths=0.5)\n",
    "    ax.set_title(metric)\n",
    "    # Set x-axis labels to classifier names\n",
    "    ax.set_xticks(range(1, len(results)+1))\n",
    "    ax.set_xticklabels(results.keys(), rotation=45)\n",
    "    \n",
    "    # Set y-axis limits\n",
    "    all_data = [item for sublist in data for item in sublist]  # Flatten the list\n",
    "    ax.set_ylim(*get_ylim(all_data))\n",
    "\n",
    "    \n",
    "# Display boxplots for each metric for all classifiers\n",
    "fig, axs = plt.subplots(2, 2, figsize=(7, 7))\n",
    "metrics = ['Precision', 'Recall', 'F1', 'FPR']\n",
    "for i, metric in enumerate(metrics):\n",
    "    plot_metric(metric, results_filtered, axs[i//2, i%2])\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"classifier_comparison_boosted.png\")\n",
    "plt.savefig(\"classifier_comparison_boosted.pdf\", format=\"pdf\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "847cb8b5",
   "metadata": {},
   "source": [
    "## Boxplots for all classifiers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "c03348fe",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x700 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "results_display = {shorthands[classifier]: results[classifier] for classifier in results}\n",
    "\n",
    "# Function to adjust the y-axis limits\n",
    "def get_ylim(data_list):\n",
    "    data_min = np.min(data_list)\n",
    "    data_max = np.max(data_list)\n",
    "    \n",
    "    # Inflate the range slightly to make sure all data points are comfortably within the plot\n",
    "    padding = (data_max - data_min) * 0.05\n",
    "    lower = data_min - padding\n",
    "    upper = data_max + padding\n",
    "    \n",
    "    # Ensure that the lower limit is not less than 0 since these metrics cannot be negative\n",
    "    lower = max(lower, 0)\n",
    "    return (lower, upper)\n",
    "\n",
    "\n",
    "# Adjust the plotting function to display results for all classifiers\n",
    "def plot_metric(metric, results_display, ax):\n",
    "    data = [results_display[classifier][metric] for classifier in results_display]\n",
    "    ax.boxplot(data, positions=range(1, len(results_display)+1), widths=0.5)\n",
    "    ax.set_title(metric)\n",
    "    # Set x-axis labels to classifier names\n",
    "    ax.set_xticks(range(1, len(results_display)+1))\n",
    "    ax.set_xticklabels(results_display.keys(), rotation=45)\n",
    "    \n",
    "    # Set y-axis limits\n",
    "    all_data = [item for sublist in data for item in sublist]  # Flatten the list\n",
    "    ax.set_ylim(*get_ylim(all_data))\n",
    "    \n",
    "    \n",
    "# Display boxplots for each metric for all classifiers\n",
    "fig, axs = plt.subplots(2, 2, figsize=(7, 7))\n",
    "metrics = ['Precision', 'Recall', 'F1', 'FPR']\n",
    "for i, metric in enumerate(metrics):\n",
    "    plot_metric(metric, results_display, axs[i//2, i%2])\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"classifier_comparison_all.png\")\n",
    "plt.savefig(\"classifier_comparison_all.pdf\", format=\"pdf\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af1a720b",
   "metadata": {},
   "source": [
    "## Saved results\n",
    "\n",
    "### 20% of the dataset, 2 runs each\n",
    "{'Logistic Regression': {'Accuracy': [0.9643332446100612, 0.964399787064147],\n",
    "  'Precision': [0.9070018924033523, 0.9077131258457375],\n",
    "  'Recall': [0.8216997305902523, 0.8214548126377663],\n",
    "  'F1': [0.8622462092007196, 0.8624325019285163],\n",
    "  'FPR': [0.013244523158664766, 0.01312901859623455]},\n",
    " 'DecisionTree': {'Accuracy': [0.9767766835240884, 0.9768099547511312],\n",
    "  'Precision': [0.9338631120225583, 0.9349948612538541],\n",
    "  'Recall': [0.8922361009061964, 0.8912564290962528],\n",
    "  'F1': [0.9125751503006012, 0.9126018808777431],\n",
    "  'FPR': [0.009933392368998576, 0.009740884764948216]},\n",
    " 'RandomForest': {'Accuracy': [0.9792387543252595, 0.979138940644131],\n",
    "  'Precision': [0.931189229618549, 0.9315684315684316],\n",
    "  'Recall': [0.9147685525349009, 0.9135439627724712],\n",
    "  'F1': [0.9229058561897703, 0.9224681587733399],\n",
    "  'FPR': [0.010626419743579871, 0.010549416701959727]},\n",
    " 'SVM': {'Accuracy': [0.9833643864785734, 0.9831980303433591],\n",
    "  'Precision': [0.9538890296427667, 0.9536004056795132],\n",
    "  'Recall': [0.9221160911094783, 0.9211364192995346],\n",
    "  'F1': [0.937733499377335, 0.9370873302603713],\n",
    "  'FPR': [0.007007276787433104, 0.007045778308243176]},\n",
    " 'AdaBoost': {'Accuracy': [0.9860593558690445, 0.985926270960873],\n",
    "  'Precision': [0.9557213930348258, 0.9545454545454546],\n",
    "  'Recall': [0.940974773450894, 0.9412196914033799],\n",
    "  'F1': [0.9482907565099346, 0.9478357380688125],\n",
    "  'FPR': [0.006853270704192816, 0.007045778308243176]},\n",
    " 'XGBoost': {'Accuracy': [0.989353207346287, 0.9896526483896726],\n",
    "  'Precision': [0.9820650781450166, 0.9818599897802759],\n",
    "  'Recall': [0.9387705118785207, 0.9412196914033799],\n",
    "  'F1': [0.9599298772852491, 0.9611104164061524],\n",
    "  'FPR': [0.0026951064567050397, 0.0027336079775151117]},\n",
    " 'LightGBM': {'Accuracy': [0.9904178866116582, 0.9906507852009582],\n",
    "  'Precision': [0.9822109275730623, 0.9834689725330621],\n",
    "  'Recall': [0.94660788635807, 0.9470977222630419],\n",
    "  'F1': [0.9640808181591419, 0.9649407361197755],\n",
    "  'FPR': [0.0026951064567050397, 0.00250259885265468]}}\n",
    "\n",
    "### 20% of the dataset, 5 runs each\n",
    "{'Logistic Regression': {'Accuracy': [0.9642928439292511,\n",
    "   0.9642928439292511,\n",
    "   0.9643460786009929,\n",
    "   0.964226300589574,\n",
    "   0.9642396092575094],\n",
    "  'Precision': [0.9002855028021571,\n",
    "   0.9005396254364617,\n",
    "   0.9008360673087099,\n",
    "   0.9002327549724926,\n",
    "   0.9001586462189318],\n",
    "  'Recall': [0.8303101228788765,\n",
    "   0.8300175541252194,\n",
    "   0.8301150770431052,\n",
    "   0.829822508289448,\n",
    "   0.8300175541252194],\n",
    "  'F1': [0.8638831109532747,\n",
    "   0.8638416645521441,\n",
    "   0.8640308582449373,\n",
    "   0.8635948442098853,\n",
    "   0.8636663453244711],\n",
    "  'FPR': [0.014533405255451954,\n",
    "   0.014487169607767589,\n",
    "   0.014440933960083224,\n",
    "   0.014533405255451954,\n",
    "   0.014548817138013408]},\n",
    " 'SVM': {'Accuracy': [0.9867046407325091,\n",
    "   0.9866780233966382,\n",
    "   0.9865848627210902,\n",
    "   0.9867179494004445,\n",
    "   0.9866114800569611],\n",
    "  'Precision': [0.9637238200220463,\n",
    "   0.9633450175262894,\n",
    "   0.9628554265118142,\n",
    "   0.9640064166833768,\n",
    "   0.9635124298315958],\n",
    "  'Recall': [0.9378779013068071,\n",
    "   0.9380729471425785,\n",
    "   0.9378779013068071,\n",
    "   0.9376828554710357,\n",
    "   0.9373902867173786],\n",
    "  'F1': [0.9506252162309099,\n",
    "   0.9505410346360987,\n",
    "   0.9502025491552218,\n",
    "   0.9506624480917539,\n",
    "   0.9502718734552645],\n",
    "  'FPR': [0.005579101487246667,\n",
    "   0.005640749017492487,\n",
    "   0.0057178084302997615,\n",
    "   0.005532865839562302,\n",
    "   0.005609925252369577]},\n",
    " 'DecisionTree': {'Accuracy': [0.9815275689056282,\n",
    "   0.9815142602376928,\n",
    "   0.9815674949094345,\n",
    "   0.9814210995621448,\n",
    "   0.9815941122453054],\n",
    "  'Precision': [0.9568219291014015,\n",
    "   0.9564411492122336,\n",
    "   0.9569294178258629,\n",
    "   0.9565038136466708,\n",
    "   0.957032457496136],\n",
    "  'Recall': [0.9055002925687536,\n",
    "   0.9057928613224108,\n",
    "   0.9056953384045251,\n",
    "   0.9050126779793252,\n",
    "   0.9057928613224108],\n",
    "  'F1': [0.930453953301934,\n",
    "   0.930428249436514,\n",
    "   0.9306077458790522,\n",
    "   0.9300461014231309,\n",
    "   0.9307079513001654],\n",
    "  'FPR': [0.006457578793249596,\n",
    "   0.006519226323495415,\n",
    "   0.00644216691068814,\n",
    "   0.0065038144409339605,\n",
    "   0.006426755028126686]},\n",
    " 'RandomForest': {'Accuracy': [0.981394482226274,\n",
    "   0.9813412475545322,\n",
    "   0.9815408775735637,\n",
    "   0.9812480868789842,\n",
    "   0.9814210995621448],\n",
    "  'Precision': [0.9694656488549618,\n",
    "   0.9690546841882154,\n",
    "   0.9698993110757816,\n",
    "   0.9700286959294293,\n",
    "   0.9697708952057701],\n",
    "  'Recall': [0.8917495611468695,\n",
    "   0.8917495611468695,\n",
    "   0.8924322215720695,\n",
    "   0.8900916715428125,\n",
    "   0.8916520382289838],\n",
    "  'F1': [0.9289850655288022,\n",
    "   0.928796343321483,\n",
    "   0.9295545736197877,\n",
    "   0.9283425723439964,\n",
    "   0.9290722487552078],\n",
    "  'FPR': [0.004438622177699006,\n",
    "   0.0045002697079448255,\n",
    "   0.004376974647453186,\n",
    "   0.0043461508823302766,\n",
    "   0.004392386530014641]},\n",
    " 'AdaBoost': {'Accuracy': [0.9887142495907585, 0.9887142495907585],\n",
    "  'Precision': [0.9682397451214656, 0.9682397451214656],\n",
    "  'Recall': [0.948410376438463, 0.948410376438463],\n",
    "  'F1': [0.9582224849738891, 0.9582224849738891],\n",
    "  'FPR': [0.004916390537104107, 0.004916390537104107]},\n",
    " 'XGBoost': {'Accuracy': [0.9919482558990671,\n",
    "   0.9920014905708088,\n",
    "   0.9921878119219047,\n",
    "   0.9920680339104859,\n",
    "   0.9922410465936464],\n",
    "  'Precision': [0.9830780014018223,\n",
    "   0.9835687806832982,\n",
    "   0.9844642678159767,\n",
    "   0.9838677354709419,\n",
    "   0.9852483692925238],\n",
    "  'Recall': [0.9574800078018334,\n",
    "   0.9573824848839477,\n",
    "   0.9578700994733762,\n",
    "   0.9575775307197192,\n",
    "   0.9574800078018334],\n",
    "  'F1': [0.970110172422311,\n",
    "   0.9702989869038794,\n",
    "   0.9709851218427166,\n",
    "   0.970544627854107,\n",
    "   0.9711657351995646],\n",
    "  'FPR': [0.002604608152885875,\n",
    "   0.0025275487400786005,\n",
    "   0.002388841797025507,\n",
    "   0.002481313092394236,\n",
    "   0.0022655467365338675]},\n",
    " 'LightGBM': {'Accuracy': [0.9927334673072572,\n",
    "   0.9926935413034509,\n",
    "   0.99266692396758,\n",
    "   0.9926403066317092,\n",
    "   0.9925471459561612],\n",
    "  'Precision': [0.9857886309047238,\n",
    "   0.9856871184065659,\n",
    "   0.9862684173599279,\n",
    "   0.985681385801542,\n",
    "   0.9858660785886126],\n",
    "  'Recall': [0.9606007411741759,\n",
    "   0.9604056953384045,\n",
    "   0.9596255119953189,\n",
    "   0.9600156036668617,\n",
    "   0.9591378974058904],\n",
    "  'F1': [0.9730317099674011,\n",
    "   0.9728821931341072,\n",
    "   0.9727645692254461,\n",
    "   0.972679215453782,\n",
    "   0.9723183391003459],\n",
    "  'FPR': [0.002188487323726593,\n",
    "   0.002203899206288048,\n",
    "   0.0021114279109193186,\n",
    "   0.002203899206288048,\n",
    "   0.0021730754411651383]}}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9553d4d8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9576d09c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}