This repository contain the codes used to obtain result tables and plots in our paper titled "Explaining the Power of Topological Data Analysis in Graph Machine Learning". Here is a description of the folders:
(1) bettiComputation folder contain codes to obtain betti distribution plots, compute betti numbers, and obtain shapely plots.
(2) filtration folder has all codes to compute Vietoris-Rips and AlphaComplex filtrations, together with their graph statistics variants.
(3) kernelAndGstats has codes for computing Weisfeiler-Leman kernel, Mapper and Graph statistics features.
(4) landscapeSilhouette has codes for the persistent landscape and silhouette feature computation.
(5) randomforest contain the code for training the randomforest model on the various feature data we have obtained.
(6) robustFiltration has the code for the robustness experiment.