This repository contains implementations of tensor methods for general (nonuniform) hypergraphs, including Z- and H-eigenvector centrality and CP decomposition from this paper.

If you use this code, please cite

@article{aksoy2024scalable, title={Scalable tensor methods for nonuniform hypergraphs}, author={Aksoy, Sinan G and Amburg, Ilya and Young, Stephen J}, journal={SIAM Journal on Mathematics of Data Science}, volume={6}, number={2}, pages={481--503}, year={2024}, publisher={SIAM} }

The hgx, hnx, simplehypergraphs, and xgi folders contain code that will be embeded in the respective hypergraph analytics libraries (HypergraphX (hgx), HyperNetX (hnx), SimpleHypergraphs.jl (simplehypergraphs), and CompleXGroupInteractions (xgi)).

The standalone-code folder contains code that you could run independent of the above libraries. The tensor-methods-for-hypergraphs-tutorial.ipynb notebook walks you through how to use the tensor times same vector functions (ttsv1, ttsv2), perform Z- and H-eigenvector centrality computations, and compute a hypergraph's associated tensor CP-decomposition.