niitsuma/sparsecorrespondenceanalysis

Correspondence Analysis optimized for sparse matrix

This project move to https://github.com/niitsuma/delayedsparse

sparse-correspondence-analysis

Correspondence Analysis(CA) is principal component analysis (PCA) of categorical data.

This project provides CA optimized for sparse matrix.

To compare existing method(Orange lib) with this Sparse CA, you can execute demo.sh.

bash  demo.sh

You can find more general description about CA in https://github.com/MaxHalford/prince#correspondence-analysis-ca

License

Author: Hirotaka Niitsuma.

@2018 Hirotaka Niirtsuma.

You can use these codes olny for self evaluation. Cannot use these codes for commercial and academical use.

pantent pending

https://patentscope2.wipo.int/search/ja/detail.jsf?docId=JP225380312

Japan patent office:patent number 2017-007741

Requirements

pip install -U sklearn Orange

In order to evaluate memory requirement, you need insall /usr/bin/time