Tensor SimRank for Heterogeneous Information Networks

TL;DR

This paper introduces Tensor SimRank, a generalized similarity measure for heterogeneous information networks that assesses object similarity based on related objects and their relations.

Contribution

It presents a novel tensor-based extension of SimRank for better similarity measurement in complex heterogeneous networks.

Findings

Effective in capturing complex relations in heterogeneous networks

Improves similarity assessment accuracy over traditional methods

Applicable to various types of information networks

Abstract

We propose a generalization of SimRank similarity measure for heterogeneous information networks. Given the information network, the intraclass similarity score s(a, b) is high if the set of objects that are related with a and the set of objects that are related with b are pair-wise similar according to all imposed relations.