Mutual Information Feature Selection (Apr 2020 – Jul 2020) is a fast, efficient open-source feature selection algorithm using mutual information.
It targets discrete–continuous datasets. Pick a dataset, choose how many top features to keep, and the algorithm returns a decision matrix, selected features, and the knobs needed to trade precision for speed.
- Mutual-information feature ranking
- Designed for discrete–continuous mixtures
- Returns decision matrix and selected features
References
B.C. Ross, “Mutual Information between Discrete and Continuous Data Sets”. PLoS ONE 9(2), 2014.
A. Kraskov, H. Stögbauer, and P. Grassberger, “Estimating Mutual Information”. Phys. Rev. E 69, 2003.
W. Gao, S. Kannan, S. Oh, P. Viswanath, “Estimating Mutual Information for discrete-continuous mixtures”. arXiv:1709.06212, 2017.
