机器学习实践之三排序和CTR预估问题

display

import pandas as pd#初始化阶段train_filename="train_small.csv"test_filename="test.csv"submission_filename="submit.csv"training_set=pd.read_csv(train_filename)

training_set.head(10)

training_set.describe()

from sklearn.externals import joblibfrom sklearn.cross_validation import train_test_splitfrom sklearn.linear_model import LogisticRegressionfrom sklearnimport metricsfrom utils import load_df%自己写的函数

在排序的场景下最常用的是评判标准是AUC
实际用得看比赛的指标