mirror of
https://github.com/Brandon-Rozek/website.git
synced 2024-12-01 18:11:08 -05:00
23 lines
472 B
Markdown
23 lines
472 B
Markdown
|
---
|
||
|
|
title: "Reverse One-Hot Encode"
|
||
|
|
date: 2020-10-11T21:58:47-04:00
|
||
|
|
draft: false
|
||
|
|
tags: []
|
||
|
|
---
|
||
|
|
|
||
|
|
Let's say that you have a dataset that is one hot encoded like the following observation:
|
||
|
|
|
||
|
|
```python
|
||
|
|
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.
|
||
|
|
|
||
|
|
```python
|
||
|
|
reverse_encoding = np.argmax(obs)
|
||
|
|
# 13
|
||
|
|
```
|
||
|
|
|