1. # ML之FE：对pandas的dataframe中的类别型字段进行数字编码化(类别型特征数值化)并导出映射表
2. from sklearn.preprocessing import LabelEncoder
3. from NDataScience.DataAnalysis import Dict2DfByRow
4. for col in df.columns:
5. print(col,df[col].dtype)
6. if df[col].dtype in ['float64', 'int', 'int64']:
7.         df[col] = pd.to_numeric(df[col])
8. else:
9. # 全部字符串化
10.         df[col] = df[col].apply(str)
11. 
12. # 编码化
13.         LbE = LabelEncoder()
14.         LbE.fit(df[col])
15.         df[col] = LbE.transform(df[col])
16. 
17. #导出映射表：开发逻辑中需要
18.         mapping_dict = dict(zip(LbE.classes_, range(1, len(LbE.classes_) + 1)))
19.         mapping_dict = {encode: label for label, encode in enumerate(LbE.classes_)}
20.         Dict2DfByRow(mapping_dict,mark=col)