Threshold-Based Heuristics for Trust Inference in a Social Network

TL;DR

This paper introduces two threshold-based heuristics for trust inference in social networks, improving computational efficiency while maintaining recommendation accuracy, especially useful for cold start users.

Contribution

It proposes novel threshold-based heuristics for trust inference that reduce computation and preserve network density and recommendation quality.

Findings

Recover up to 70% of trust paths with less computation

Heuristics maintain recommendation accuracy

Inferred networks reflect trust propagation phenomena

Abstract

Trust among the users of a social network plays a pivotal role in item recommendation, particularly for the cold start users. Due to the sparse nature of these networks, trust information between any two users may not be always available. To infer the missing trust values, one well-known approach is path based trust estimation, which suggests a user to believe all of its neighbors in the network. In this context, we propose two threshold-based heuristics to overcome the limitation of computation for the path based trust inference. It uses the propagation phenomena of trust and decides a threshold value to select a subset of users for trust propagation. While the first heuristic creates the inferred network considering only the subset of users, the second one is able to preserve the density of the inferred network coming from all users selection. We implement the heuristics and analyze the inferred networks with two real-world datasets. We observe that the proposed threshold based heuristic can recover up to 70 \% of the paths with much less time compared to its deterministic counterpart. We also show that the heuristic based inferred trust is capable of preserving the recommendation accuracy.