Skip to content

[Research] Benchmark clustering E step + M steps #25

Description

@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.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    researchResearching new methods

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions