[Research] Benchmark clustering E step + M steps

[iraedeus]

This issue refers to this PR.

Conduct an experiment using an experimental environment to determine the applicability of the clusterization in step E.

Experiment setup

Distributions: Exponential, normal and Weibull
Algorithms: Clustering E step + LikelihoodMStep against BayesianEStep + LikelihoodMStep
The remaining parameters of the experiment and hyperparameters of the classifiers are selected by the performer.

The expected result is a table of statistics by metric, as well as an estimate of the algorithm speed. Ideally, it is also necessary to select ideal hyperparameters or at least study their influence on the result.