ok

demo/kaggle-higgs/README.md

@@ -0,0 +1 @@
+This is the folder giving example of how to use XGBoost to run Kaggle Higgs competition

demo/kaggle-higgs/higgs-numpy.py

@@ -0,0 +1,50 @@
+#!/usr/bin/python
+# this is the example script to use xgboost to train 
+import sys
+import numpy as np
+# add path of xgboost python module
+sys.path.append('../../python/')
+import xgboost as xgb
+
+test_size = 550000
+
+# path to where the data lies
+dpath = 'data'
+
+# load in training data, directly use numpy
+dtrain = np.loadtxt( dpath+'/training.csv', delimiter=',', skiprows=1, converters={32: lambda x:int(x=='s') } )
+label  = dtrain[:,32]
+data   = dtrain[:,1:31]
+# rescale weight to make it same as test set
+weight = dtrain[:,31] * float(test_size) / len(label)
+
+sum_wpos = sum( weight[i] for i in xrange(len(label)) if label[i] == 1.0  )
+sum_wneg = sum( weight[i] for i in xrange(len(label)) if label[i] == 0.0  )
+
+# print weight statistics 
+print 'weight statistics: wpos=%g, wneg=%g, ratio=%g' % ( sum_wpos, sum_wneg, sum_wneg/sum_wpos )
+
+# construct xgboost.DMatrix from numpy array, treat -999.0 as missing value
+xtrain = xgb.DMatrix( data, label=label, missing = -999.0 )
+
+# setup parameters for xgboost
+params = {}
+# use logistic regression loss
+param['loss_type'] = 3
+# scale weight of positive examples
+param['scale_pos_weight'] = sum_wpos/sum_wpos
+param['bst:eta'] = 0.1 
+param['bst:max_depth'] = 6
+param['eval_metric'] = '[email protected]'
+param['silent'] = 1
+param['eval_train'] = 1
+param['nthread'] = 16
+
+# boost 120 tres
+num_round = 120
+print 'loading data end, start to boost trees'
+bst = xgb.train( xtrain, param, num_round );
+# save out model
+bst.save_model('higgs.model')
+
+print 'finish training'

demo/kaggle-higgs/higgs-pred.py

@@ -0,0 +1,52 @@
+#!/usr/bin/python
+# this is the example script to use xgboost to train 
+import sys
+import numpy as np
+# add path of xgboost python module
+sys.path.append('../../python/')
+import xgboost as xgb
+
+# path to where the data lies
+dpath = 'data'
+
+modelfile = 'higgs.model'
+outfile = 'higgs.pred.csv'
+# make top 15% as positive 
+threshold_ratio = 0.15
+
+# load in training data, directly use numpy
+dtest = np.loadtxt( dpath+'/test.csv', delimiter=',', skiprows=1 )
+data   = dtest[:,1:31]
+idx = dtest[:,1]
+
+xtest = xgb.DMatrix( data, missing = -999.0 )
+bst = xgb.Booster()
+bst.load_model( modelfile )
+
+ypred = bst.predict( dtest )
+res  = [ ( int(idx[i]), ypred[i] ) for i in xrange(len(ypred)) ] 
+
+rorder = {}
+for k, v in sorted( res, key = lambda x:-x[1] ):
+    rorder[ k ] = len(rorder) + 1
+
+# write out predictions
+ntop = int( ratio * len(rorder ) )
+fo = open(outfile, 'w')
+nhit = 0
+ntot = 0
+fo.write('EventId,RankOrder,Class\n')
+for k, v in res:        
+    if rorder[k] <= ntop:
+        lb = 's'
+        nhit += 1
+    else:
+        lb = 'b'        
+    fo.write('%s,%d,%s\n' % ( k, rorder[k], lb ) )
+    ntot += 1
+fo.close()
+
+print 'finished writing into model file'
+
+
+

demo/kaggle-higgs/run.sh

@@ -0,0 +1,4 @@
+#!/bin/bash
+
+./higgs-numpy.py
+./higgs-pred.py