CODEWITHSUNDEEP
python
Jane Borges
Jane Borges

Reputation: 592

Application function LabelEncoder().fit_transform in python

I have a theoretical question about the function LabelEncoder().fit_transform.

I am using the function / method in a classification application. It's working perfectly.

       #Import
       from sklearn.preprocessing import LabelEncoder

       #Transform original values ​​by encoded labels
       df_data = df_data.apply(LabelEncoder().fit_transform)

However, in the documentation "sklearn.preprocessing.LabelEncoder" have: "This transformer should be used to encode target values, i.e. y, and not the input X".

I am applying this method across the dataframe. Numeric input variables (X) and categorical output variable (y). I thought of applying in X to transform the objective variable into the numeric type and I thought of applying in y to deal with problems of magnitude between different sources of input variables. Is this attitude correct? Is there another function that I can apply in place of LabelEncoder().fit_transform for input variables? Thank you

Upvotes: 0

Views: 677

Answers (1)

gnodab
gnodab

Reputation: 880

As the documentation states Label Encoder should only be used to transform your lables. i.e from 'Apple', 'Orange' to 0, 1. If you have categorical labels then look at the One Hot Encoder. Additionally, if your input X have differing scales, then take a look at the Standard Scaler

Upvotes: 1

Related Questions