电信用户匹配模型实例

如下：

 

import numpy as np

import pandas as pd

import jieba.analyse as jieba

 

data = pd.read_csv(r'/opt/dianxing/train_all.csv',low_memory=False)

def dataProcess(data):

data.is_mix_service=data.is_mix_service.fillna(data.is_mix_service.mode()[0])

return data

Train_data = dataProcess(train_data)

feature = ['service_type’,’is_mix_service’,’online_time’,’1_total_fee’,’2_total_fee’,’3_total_fee’,’4_total_fee’,’month_traffic’,’many_over_bill’,’contract_type’,’contract_time’,’is_promise_low_consume’,’net_service’,’pay_times’,’pay_num’,’last_month_traffic’,’local_trafffic_month’,’local_caller_time’,’service1_caller_time’,’gender’,’age’,’complaint_level’,’former_complaint_num’,’former_complaint_fee’,’current_service’]

feature = ['service_type','is_mix_service','month_traffic','pay_times','pay_num','former_complaint_num','current_service','user_id']

x = data[feature]

y = data.service_type

from sklearn.tree import DecisionTreeClassifier as DT

clf = DT()

clf = fit(x,y)

 

from sklearn.svm import SVC

clf = SVC()

clf.fit(x, y)