https://github.com/dholzmueller/realmlp-td-s_standalone

Standalone implementation of RealMLP-TD-S for tabular data classification and regression
https://github.com/dholzmueller/realmlp-td-s_standalone

Last synced: 24 days ago
JSON representation

Standalone implementation of RealMLP-TD-S for tabular data classification and regression

• Host: GitHub
• URL: https://github.com/dholzmueller/realmlp-td-s_standalone
• Owner: dholzmueller
• License: mit
• Created: 2024-07-08T09:23:37.000Z (6 months ago)
• Default Branch: main
• Last Pushed: 2024-07-08T09:26:20.000Z (6 months ago)
• Last Synced: 2024-07-08T11:49:48.458Z (6 months ago)
• Language: Python
• Size: 5.86 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.txt

Awesome Lists containing this project

README

        
# RealMLP-TD-S standalone implementation

This repository contains a small standalone implementation of RealMLP-TD-S, 

a neural network for tabular datasets, consisting of

- preprocessing code (91 lines of code) in `preprocessing.py`, which includes 

  - one-hot encoding with custom missing/unknown value encoding 

and encoding binary categories to 1/-1

  - robust scaling and smooth clipping

- the MLP implementation (211 lines of code) in `mlp.py`.

The implementation is standalone in the sense that it only uses 

`numpy`, `pandas`, `sklearn`, and `torch`. 

This code is not available on `pip` since it can just be copied.

RealMLP-TD-S is also available with more functionality 

in [PyTabKit](https://github.com/dholzmueller/pytabkit).

The file `check_mlp.py` 

checks that this implementation matches the one in PyTabKit

and requires to install PyTabKit to run.