Cannot find the implementation for paper MIND about kxy-python HOT 2 CLOSED

Hello LJSthu,

Here is an implementation of MIND under Spearman rank correlation constraints that you may use to reproduce experiments in the paper: https://github.com/kxytechnologies/kxy-python/blob/master/kxy/misc/mind.py.

The kxy backend utilizes a more advanced mutual information estimator; it supports ordinal (discrete and continuous) and non-ordinal variables, and the maximum-entropy constraints are learned. To retrieve the mutual information estimated by the backend, call one of the two data_valuation functions with include_mutual_information set to True (see test_data_valuation.py for an example).

Yves-Laurent

Yves-Laurent

from kxy-python.

Hello drylks,

If the numbers of samples in the two distributions x and y are different, e.g. 500 samples for x and 1000 samples for y, is it correct to broadcast x to the length of y when calculating mutual information?

from kxy-python.