Random Walks on Simplicial Complexes and the normalized Hodge Laplacian

Modeling complex systems and data with graphs has been a mainstay of the applied mathematics community. The nodes in the graph represent entities and the edges model the relations between them. Simplicial complexes, a mathematical object common in topological data analysis, have emerged as a model for multi-nodal interactions that occur in several complex systems; for example, biological interactions occur between a set of molecules rather than just two, and communication systems can have group messages and not just person-to-person messages. While simplicial complexes can model multi-nodal interactions, many ideas from network analysis concerning dynamical processes lack a proper correspondence in the a simplicial complex model. In particular, diffusion processes and random walks, which underpin large parts of the network analysis toolbox including centrality measures and ranking, community detection, and contagion models, have so far been only scarcely studied for simplicial complexes. Here we develop a diffusion process on simplicial com- plexes that can serve as a foundation for making simplicial complex models more effective. Our key idea is to generalize the well-known relationship between the normalized graph Laplacian operator and random walks on graphs by devising an appropriate normalization for the Hodge Laplacian, the analog of the graph Laplacian for simplicial complexes. We demonstrate how our diffusion process can be used for system analysis by developing a generalization of PageRank for edges in simplicial complexes and analyzing a book co-purchasing dataset.