Skip to content

BayesianGaussianMixture: Wrong documentation on mean_precision_prior #12273

@gmeisl

Description

@gmeisl

Description

According to the documentation:
mean_precision_prior :
... Smaller values concentrate the means of each clusters around mean_prior. ...

However, for me with a large precision prior the mean_prior is preserved and for low values the the posterior mean shows stronger variation.

Steps/Code to Reproduce

Try on data of your choice, set a mean_prior and compare results for large and small mean_precision_prior (1e-10, 1e+10).

Versions

Windows-10-10.0.15063-SP0
Python 3.6.5 |Anaconda, Inc.| (default, Mar 29 2018, 13:32:41) [MSC v.1900 64 bit (AMD64)]
NumPy 1.14.3
SciPy 1.1.0
Scikit-Learn 0.19.1

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions