【sklearn】sklearn.preprocessing.LabelEncoder

sklearn.preprocessing.LabelEncoder

effect ： Convert non numeric labels to numeric labels

---

First, let's see what the official said ： Official address

• Encode target labels with value between 0 and n_classes-1.
• take Value of target label Encoded as 1 To n-1, （ altogether n A classification ）
• This transformer should be used to encode target values, i.e. y, and not the input X.
• This transformation Should be applied to code Value of label , such as y, Instead of encoding input values , such as x

---

give an example

import numpy as np
import pandas as pd
import sklearn.preprocessing as preprocessing

city = pd.Series(["paris", "paris", "tokyo", "amsterdam"])

le = preprocessing.LabelEncoder()
le.fit(city)
list(le.classes_) #  Check the category after label weight removal 
# ['amsterdam', 'paris', 'tokyo']

#  Tag value conversion , It can be understood as , Map non numeric labels to numeric labels 
le.transform(city)
#  Output is ： array([1, 1, 2, 0])

#  Reverse transformation , Restore digital labels to non digital labels 
list(le.inverse_transform([1, 1, 2, 0]))
# ['tokyo', 'tokyo', 'paris']

What is involved above 3 A way ：

• le.fit(df) : Remove the weight of the label , Generate categories , And encoded as a digital label
• le.transform(df) ： Map all the original data labels to On the digital label
• le.inverse_transform(list) ： Put the number tag Inverse conversion to Non numeric label