Link prediction with continuous-time classical and quantum walks

TL;DR

This paper introduces a new link prediction method for protein-protein interaction networks using continuous-time classical and quantum random walks, demonstrating competitive performance on real datasets.

Contribution

It proposes a novel class of link prediction techniques based on continuous-time classical and quantum walks, utilizing network matrices to improve prediction accuracy.

Findings

Classical random walks effectively predict missing interactions.

Quantum walks with adjacency matrix perform well in link prediction.

Method rivals state-of-the-art approaches on real-world datasets.

Abstract

Protein-protein interaction (PPI) networks consist of the physical and/or functional interactions between the proteins of an organism. Since the biophysical and high-throughput methods used to form PPI networks are expensive, time-consuming, and often contain inaccuracies, the resulting networks are usually incomplete. In order to infer missing interactions in these networks, we propose a novel class of link prediction methods based on continuous-time classical and quantum random walks. In the case of quantum walks, we examine the usage of both the network adjacency and Laplacian matrices for controlling the walk dynamics. We define a score function based on the corresponding transition probabilities and perform tests on four real-world PPI datasets. Our results show that continuous-time classical random walks and quantum walks using the network adjacency matrix can successfully predict missing protein-protein interactions, with performance rivalling the state of the art.