This project implements a selection of fundamental algorithms used in topological data analysis. The purpose is educational, for most use cases there are more optimised and tested libraries. I have aimed to keep the implementations simple and compact to highlight the algorithms. The library is written in python only, and relies on numpy, scipy and scikit-learn for the heavy lifting.
The implemented algorithms include
- Graph learning algorithms
- Self Organising Maps
- Growing Neural Gas
- Generative Gaussian Graphs
- Mapper
- Persistent homology
- Vietoris-Rips Complex
- Alpha Complex
- Matrix reduction for finding birth-death pairs
- Persistence images
- Persistence landscapes
- Toplogical loss
- Topological loss function for autoencoders
See the example notebooks for more information.