compute_model_covariances_weights (
        dataset,
        labels,
        means,
        )

Estimates the the covariances of each class given a dataset, class labeling and class means. The results are returned an a three dimensional Numeric array.

compute_model_means ( dataset,  labels )

Given a dataset and a labeling compute the means for each cluster and return the results as a two dimensional Numeric array with each row corresponding to a particular class.

compute_model_weights ( dataset,  labels )

Evaluates the weight of each class realtive to the whole. This is simply this number of datapoints in each class divided by the total number of datapoints.

constructMixtureOfDiagonalGaussiansFromLabeling ( dataset,  labels )

Return an estimated mixture of diagonal gaussians class instance from a dataset and labeling.

constructMixtureOfFullGaussiansFromLabeling ( dataset,  labels )

Return an estimated mixture of full gaussians class instance from a dataset and labeling.

constructMixtureOfGaussiansFromLabeling ( dataset,  labels )

Return an estimated mixture of gaussians class instance from a dataset and labeling.

estimateParameters ( dataset,  labels )

Fully estimate the Mixture of Gaussians parameters for a dataset with a given hard partitioning. The results are returned as a 4-tuple: (k, means, covariances, weights).

See the documentation for compute_model_means() and compute_model_variances() for details of returned data.