One-Hot Encoding

Categorical data is very common, for example occupation would fall into several categories rather than a numerical value. This can be problematic for many machine learning algorithms. Instead a common approach is to convert them into numerical values via One-Hot Encoding. It works best when a categorical variable only takes on a small number of values.

One-Hot encoding replaces a column with new binary columns that indicate the presence of each possible value. For example:

This can be done in pandas with the method get_dummies. For example the Python code

import pandas as pd
 
df = pd.DataFrame({
    'Sex': ['Male', 'Female', 'Female', 'Male', 'Male'],
    'Color': ['Red', 'Red', 'Yellow', 'Green', 'Yellow']
})
df_hot_encoded = pd.get_dummies(df)

produces the following table:

Sex_Female	Sex_Male	Color_Green	Color_Red	Color_Yellow
0	1	0	1	0
1	0	0	1	0
1	0	0	0	1
0	1	1	0	0
0	1	0	0	1