website/content/blog/reverseonehotencode.md

title	date	draft	tags	medium_enabled
Reverse One-Hot Encode	2020-10-11T21:58:47-04:00	false	Python	true

Let's say that you have a dataset that is one hot encoded like the following observation:

import numpy as np
obs = np.array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0])

The easiest way to reverse one-hot encode the structure, is to take the argmax of the observation.

reverse_encoding = np.argmax(obs)
# 13