Exercise1 Yiran Li

Exercise1/LR_breastcancer.py

@@ -0,0 +1,51 @@
+from sklearn.model_selection import train_test_split
+from sklearn.linear_model import LinearRegression,SGDRegressor,Ridge,LogisticRegression
+from sklearn.metrics import mean_squared_error, classification_report
+from sklearn.preprocessing import StandardScaler
+from sklearn.metrics import mean_squared_error
+import pandas as pd
+import numpy as np
+
+def logistic():
+    """
+    逻辑回归做二分类进行乳腺癌预测，用数据集的0.7作为训练集，用数据集的0.3作为测试集。Pytorch进行二分类要分每个维度分别进行，这里有sklearn内置逻辑回归函数对数据进行分类。首先读取在线数据集，然后进行数据分割，按0.7/0.3分割为训练集和测试集。
+    """
+    # 构造列标签名字
+    column = ['Sample code number', 'Clump Thickness', 'Uniformity of Cell Size', 'Uniformity of Cell Shape',
+              'Marginal Adhesion', 'Single Epithelial Cell Size', 'Bare Nuclei', 'Bland Chromatin', 'Normal Nucleoli',
+              'Mitoses', 'Class']
+    # 读取在线数据集
+    data = pd.read_csv(
+        "https://archive.ics.uci.edu/ml/machine-learning-databases/breast-cancer-wisconsin/breast-cancer-wisconsin.data",
+        names=column)
+    # print(data)
+    # 对存在缺失值进行处理，替换为nan
+    data = data.replace(to_replace='?', value=np.nan)
+    data = data.dropna()
+    # 输出data的数据量和维度。
+    print("数据量，维度：\n", data.shape)
+    print("数据集：\n", data)
+    # 进行数据的分割，用数据集的0.7作为训练集，用数据集的0.3作为测试集
+    x_train, x_test, y_train, y_test = train_test_split(data[column[1:10]], data[column[10]], test_size=0.3)
+    # 输出训练样本两个类分别的样本量。
+    print("训练样本类，数据量：\n", y_train.value_counts())
+    # 输出测试样本两个类分别的样本量。
+    print("测试样本类，数据量：\n", y_test.value_counts())
+    # 进行标准化处理
+    std = StandardScaler()
+    x_train = std.fit_transform(x_train)
+    x_test = std.transform(x_test)
+    # 逻辑回归来对训练集进行训练
+    lg = LogisticRegression(C=1.0)
+    # 求得训练集X的均值、方差、最大值、最小值等固有属性
+    lg.fit(x_train, y_train)
+    print('回归系数\n', lg.coef_)
+    # 训练后返回预测结果
+    y_predict = lg.predict(x_test)
+    # 输出预测结果计算出的决定系数R^2
+    print("拟合优度：", lg.score(x_test, y_test))
+    # classification_report函数用于显示主要分类指标．显示每个类的精确度，召回率，F1值等
+    print("分类指标：", classification_report(y_test, y_predict, labels=[2, 4], target_names=["良性", "恶性"]))
+
+if __name__ == "__main__":
+        logistic()