Affinis

Tools for inferring relations from co-occurrence data

What does it help with?

Affinis is a tool for assisting in unsupervised structure learning on sparse, binary data.

In large (sparse) feature matrices, especially ones with binary or integer-valued entries, you commonly need to figure out the underlying structure of your feature space from the observations.

Techniques for this are widely varied, and different communities have widely different practices and assumptions for what is an appropriate approach.

For an overview of features and usage, see the User Guide.
For an example of using the library on a (toy) dataset, see our example. Theoretical discussion of our algorithms, and benchmarks for them, can be found among the chapters of the online mirror for Sexton (2025).