maxent_toolbox

Maximum Entropy toolbox for MATLAB

View the Project on GitHub orimaoz/maxent_toolbox

maxent.getEmpiricalModel

Description

Returns an empirical distribution corresponding to a set of samples.

Usage

model_out = maxent.getEmpiricalModel(samples)
model_out = maxent.getEmpiricalModel(samples,Name,Value,...)

Arguments

Mandatory arguments

  • samples - Set of samples to train the model on, in the format (ncells x nsamples).

Optional arguments (in the form of name,value pairs)

  • min_count - ignore all samples that appears less than this number of times (default 1, which does not ignore any samples).

Output

  • model_out - structure with several fields describing the empirical distrbution:
    • model_out.words - unique codewords in the empirical distribution.
    • model_out.counts - how many times each of these codewords repeated in the empirical dataset.
    • model_out.logprobs - log probabilities of the codewords in the empirical distribution.
    • model_out.entropy - entropy of the empirical distribution, in bits.