bmdistribution

Multivariate binary (Bernoulli) mixture model distribution and fitting

An implementation of the multivariate Binary (Bernoulli) mixture model for Matlab. Can be used to model distribution of pixels in distributions of binary images for example. This library follows the matlab distribution class as closely as possible, and more precisely the Gaussian mixture model one.
Example:
"""
Xlearn = [bsxfun(@lt, rand(40000, 9), [.1 .2 .3 .4 .5 .6 .7 .8 .9]);
bsxfun(@lt, rand(10000, 9), [.9 .8 .7 .6 .5 .4 .3 .2 .1])];
bmm = fitbmdist(Xlearn, 2, 'Options', struct('MaxIter', 1000, 'TolFun', 1e-4))
bmm.ComponentProportion
bmm.Means
Xtest = [bsxfun(@lt, rand(1000, 9), [.1 .2 .3 .4 .5 .6 .7 .8 .9]);
bsxfun(@lt, rand(1000, 9), [.9 .8 .7 .6 .5 .4 .3 .2 .1])];
clus = bmm.cluster(Xtest);
% label numbers might be inverted, check numbers and build confusion matrix.
"""