This directory contains all code used for the project. The project structure is broken down as follows:
- data: wrappers for various datasets
- distributions: utils for various distributions (deprecated)
- generation: generating simulated datasets
- hypothesis: hypothesis tests on feature-based partitions
- inference: softmax classifier and MALA
- model: includes Graph_MCMC class used to sample block memeberships
- output: directory to save output files
- sgmcmc: brief foray into R (deprecated)
- utils: various utility functions
The top level files are either scripts or jupyter notebooks.
The script opus.py contains the final experiment used in the report.