datamicroscopes: Bayesian nonparametric models in Python

datamicroscopes is a library for discovering structure in your data. It implements several Bayesian nonparametric models for clustering such as the Dirichlet Process Mixture Model (DPMM) , the Infinite Relational Model (IRM) , and the Hierarchichal Dirichlet Process (HDP) . These models rely on the Dirichlet Process, which allow for the automatic learning of the number of clusters in a datset. Additionally, our API provides users with a flexible set of likelihood models for various types of data, such as binary, ordinal, categorical, and real-valued variables( datatypes) .

Please read our introduction for an overview of clustering and structure discovery.


First, install Anaconda. Then in the terminal type:

$ conda config --add channels distributions
$ conda config --add channels datamicroscopes
$ conda install microscopes-common
$ conda install microscopes-{mixturemodel, irm, lda}
Datamicroscopes is developed by Qadium, with funding from the DARPA XDATA program. Copyright Qadium 2015.