Home

Awesome

mixed_KSG

This is the Python Code for the paper "Estimating Mutual Information for Discrete-Continuous Mixtures", NIPS 2017.

Reference: http://arxiv.org/abs/1709.06212

Sample Usage:

import mixed 
MI = mixed.mixed_KSG(x,y)

Input

X -- 2D array of size N by d_x (1D array of size N if d_x = 1)

Y -- 2D array of size N by d_y (1D array of size N if d_y = 1)

Output

An estimate of I(X;Y)

See demo.py for an example (Experiment I in Section 5 of the paper)

Contact wgao9@illinois.edu