构建决策树时出现ValueError: Length of feature_names, 4 does not match number of features, 10的解决办法

最新推荐文章于 2024-07-13 07:15:00 发布

开大卡车的

最新推荐文章于 2024-07-13 07:15:00 发布

阅读量8.1k

点赞数 18

文章标签： python 机器学习数据分析决策树

本文链接：https://blog.csdn.net/weixin_43100085/article/details/110602936

版权

在构建决策树时遇到`ValueError: Length of feature_names, 4 does not match number of features, 10`。问题源于`feature_names`参数需与数据集特征名一致，但经过`pd.get_dummies()`独热编码后，特征名数量变为10。正确做法是将`feature_names`设置为独热编码后的特征名，例如：Outlook_overcast, Outlook_rain, Outlook_sunny, Temperature_cool, ..., Windy_yes。" 131440715,14311535,Himall商城优惠券管理功能详解,"['java', '开发语言', '商城系统', '优惠活动', '积分系统']

摘要由CSDN通过智能技术生成

构建决策树时出现ValueError: Length of feature_names, 4 does not match number of features, 10的解决办法

以下为原代码：

import pandas as pd
from sklearn import tree

ball_data = pd.read_csv('ball.csv')
cat_features = ['Outlook','Temperature','Humidity','Windy']
ball_data_onehot = pd.get_dummies(ball_data,columns=cat_features)
ball_data_onehot['Play'] = ball_data_onehot['Play'].map({
   'No':0,'Yes':1}).astype(int)
ball_data_tr_in = ball_data_onehot.drop(columns=['Play'])
ball_data_tr_out = ball_data_onehot['Play']

from sklearn.model_selection import train_test_split

X_train,X_test,y_train,y_test = train_test_split(ball_data_tr_in,ball_data_tr_out,test_size=0.3,random_state=1