python模型修改

因为一门课程需要，想把之前18年的ccf大赛题目面向电信行业存量用户的智能套餐个性化匹配模型给修改使用一下，但在借鉴了几位博主的模型后各种警告和报错，现在两眼一抹黑，想问一下该怎么修改
比如这个xgb方法

import lightgbm as lgb
import pandas as pd
import numpy as np
from sklearn.preprocessing import LabelEncoder
from sklearn.model_selection import train_test_split
from sklearn.metrics import f1_score
import xgboost as xgb
train = pd.read_csv('./train.csv')
test = pd.read_csv('./test.csv')
print (train.shape)



train = train[train.gender != '\\N']
# test = test[test.gender != '\\N']
train['gender'] = train['gender'].apply(lambda x : int(x))
test['gender'] = test['gender'].apply(lambda x : int(x))

train = train[train.age != '\\N']
# test = test[test.age != '\\N']
train['age'] = train['age'].apply(lambda x : int(x))
test['age'] = test['age'].apply(lambda x : int(x))

train = train[train['2_total_fee'] != '\\N']
# test = test[test['2_total_fee'] != '\\N']
test.loc[test['2_total_fee'] == '\\N','2_total_fee'] = 0.0
train['2_total_fee'] = train['2_total_fee'].apply(lambda x : float(x))
test['2_total_fee'] = test['2_total_fee'].apply(lambda x : float(x))

train = train[train['3_total_fee'] != '\\N']
# test = test[test['3_total_fee'] != '\\N']
test.loc[test['3_total_fee'] == '\\N','3_total_fee'] = 0.0
train['3_total_fee'] = train['3_total_fee'].apply(lambda x : float(x))
test['3_total_fee'] = test['3_total_fee'].apply(lambda x : float(x))


label = train.pop('current_service')
le = LabelEncoder()
label = le.fit_transform(label)


feature = [value for value in train.columns.values if
                   value not in ['user_id']]



#xgb模型
def XGB():
    clf = xgb.XGBClassifier(max_depth=12, learning_rate=0.05,
                            n_estimators=752, silent=True,
                            objective="multi:softmax",
                            nthread=4, gamma=0,
                            max_delta_step=0, subsample=1, colsample_bytree=0.9, colsample_bylevel=0.9,
                            reg_alpha=1, reg_lambda=1, scale_pos_weight=1,
                            base_score=0.5, seed=2018, missing=None,num_class=15)
    return clf




online = False
# online = True
if online:
        print ('online')

        model = XGB()
        model.fit(train[feature], label, eval_set=[(train[feature], label)], verbose=1,)
        pred = model.predict(test[feature])
        pred = le.inverse_transform(pred)
        test['predict'] = pred

        test[['user_id', 'predict']].to_csv('./sub.csv', index=False)
else:
        print ('offline')
        train_x,test_x,train_y,test_y = train_test_split(train[feature],label,test_size=0.1,shuffle=True,random_state=2018)
        model = XGB()
        model.fit(train_x[feature], train_y, eval_set=[(test_x[feature], test_y)], verbose=1,early_stopping_rounds=100)
        pred = model.predict(test_x)
        print(f1_score(test_y,pred,average='weighted'))

        # feature_list = model.feature_importances_
        # pd.DataFrame(
        #         {
        #                 'feature':feature,
        #                 'score':feature_list,
        #         }
        # ).to_csv('./feature_importance.csv',index=False)

        # from sklearn.externals import joblib
        # joblib.dump(model, 'gbm.pkl')

我在使用课程的训练和测试集后报错

img


还有一开始的一些警告

img

数据集和测试集
[](链接：https://pan.baidu.com/s/10G9HqBzF54Ki9S7IwMrYQw
提取码：b2kx)

叙述详细点，不清楚你的要求

csv发下，好测试

你直接把代码贴出来被

CCF大数据竞赛-面向电信行业存量用户的智能套餐个性化匹配模型
可以借鉴下
https://blog.csdn.net/qq_34783311/article/details/83472609

从你提供的错误来看，有可能是你的数据集的问题，比如数据的类型。当然也不排除代码中参数的设置问题。建议对比下你的数据集和这个程序的数据要求。再调试代码

对比一下之前的功能、数据和现在的功能、数据之间哪里有区别，然后正对性的修改

第78行报空值错误，再此处打断点调试，然后看看运行到这里的数据情况

大体思路，如果你想要修改max_depth参数的值，可以将max_depth=12改为你想要的任何其他值。
下面是一个修改max_depth参数的示例：

def XGB():  
    clf = xgb.XGBClassifier(max_depth=5, learning_rate=0.05,  
                            n_estimators=752, silent=True,  
                            objective="multi:softmax",  
                            nthread=4, gamma=0,  
                            max_delta_step=0, subsample=1, colsample_bytree=0.9, colsample_bylevel=0.9,  
                            reg_alpha=1, reg_lambda=1, scale_pos_weight=1,  
                            base_score=0.5, seed=2018, missing=None,num_class=15)  
    return clf