classification_precedure_python

import sklearn
import pandas as pd
import numpy as np

'''----------导入导出数据---------'''

file1 = '/Users/e/Desktop/active_lost_prediction_model.csv'
t = pd.read_csv(file1)

#透视表
t = t.pivot_table(index = ['user_id','exist_type'],columns = 'service',aggfunc = max,fill_value = 0)

file2 = '/Users/e/Desktop/export.csv'
t.to_csv(file2)

'''----------数据处理---------'''
#填充缺失值
t_pro = t.fillna(-1)

#替换值
t_pro['sex'] = t_pro['sex'].replace(['男','女'],[1,0])

replace_dict = {
    -1.0:20,
    1.0:1,
    2.0:20
}
t_pro['sex'] = t_pro['sex'].map(lambda x:replace_dict[x])


'''----------平衡类---------'''
#拆分不同类，统计个数
t_pro_label1 = t_pro[t_pro['label']==1]
t_pro_label0 = t_pro[t_pro['label']==0]

label1_cnt = t_pro_label1.count()
label0_cnt = t_pro_label0.count()

#抽样
t_pro_label1_sampled = t_pro_label1.sample(n = label0_cnt)

#合并
t_prod = pd.concat(t_pro_label1_sampled,t_pro_label0,axis = 0)


'''----------划分训练集---------'''
#拆分X与y
y = data['label']
X = data.drop(['label','user_id','selected_day'],axis = 1)


#拆分训练集
from sklearn.cross_validation import train_test_split
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.3,random_state = 0,stratify = y)

'''----------训练与预测模型---------'''
from sklearn.ensemble import RandomForestClassifier
forest = RandomForestClassifier(criterion = 'entropy',n_estimators = 10,random_state=1,n_jobs = 2)
forest.fit(X_train,y_train)

y_pred = forest.predict(X_test)
y_train_pred = forest.predict(X_train)




'''----------训练与测试的准确性---------'''
from sklearn.metrics import classification_report
from sklearn.metrics import accuracy_score
print('predict')
print(accuracy_score(y_test,y_pred))
print(classification_report(y_test,y_pred))
print('model')
print(accuracy_score(y_train,y_train_pred))
print(classification_report(y_train,y_train_pred))




'''----------随机森林相关函数---------'''
#特征重要性
importances = forest.feature_importances_
feature_labels = X_train.columns
indices = np.argsort(importances)[::-1]
for f in range(X_train.shape[1]):
    print("%d  %20s  %f",f+1,feature_labels[indices[f]],importances[indices[f]])