encoding for class labels

1. use pandas

import numpy as np

""" coverted string class to integer """
class-map = { label : idx for idx, label in enumerate(np.unique(df['class']))}

""" reversed integer to strings """
df['class'] = df['class'].map(class-map)

 

2. use scikit-learn

from sklearn.preprocessing import LabelEncoder
enc = LabelEncoder()

""" converted integer to strings """"
y = enc.fit_transform(df['class'].values)

""" reversed integer to strings """
enc.inverse_transform(y)