mnist-classifier 1.0.4

Basic mnist classifier example of a Reproducible Research Project in Python

M-NIST classification algorithm comparison

Installation

You can just use

pip install mnist-classifier

Documentation

You can find all the information you need on the documentation page

Motivation for project

This project was realised in the scope of a course in Artificial Intelligence offered by UniDistance and the Idiap research Institute

The hypothesis motivating the development of this package is the following:

Random Forests can give similar resulting prediction models to MLP Neural Networks on the M-NIST digit dataset in significantly less time.

With the code in this repository, we show that indeed, Random Forests can in fact produce similar (if not better) results with training times orders of magnitude smaller.

License

MIT

Authors

@sandrich - Christian Sandrini @bigskapinsky - Calixte Mayoraz