Topic-based Community Search over Spatial-Social Networks (Technical Report)

TL;DR

This paper introduces a new community search problem over spatial-social networks that considers social influence, travel time, and keywords, providing efficient algorithms and indexing for practical applications.

Contribution

It formulates the TCS-SSN problem, proposes pruning techniques, a novel social-spatial index, and an efficient query algorithm for community search in spatial-social networks.

Findings

Pruning techniques significantly reduce search space.

The social-spatial index improves query efficiency.

Experimental results validate effectiveness on real and synthetic data.

Abstract

Recently, the community search problem has attracted significant attention, due to its wide spectrum of real-world applications such as event organization, friend recommendation, advertisement in e-commence, and so on. Given a query vertex, the community search problem finds dense subgraph that contains the query vertex. In social networks, users have multiple check-in locations, influence score, and profile information (keywords). Most previous studies that solve the CS problem over social networks usually neglect such information in a community. In this paper, we propose a novel problem, named community search over spatial-social networks (TCS-SSN), which retrieves community with high social influence, small traveling time, and covering certain keywords. In order to tackle the TCS-SSN problem over the spatial-social networks, we design effective pruning techniques to reduce the problem search space. We also propose an effective indexing mechanism, namely social-spatial index, to facilitate the community query, and develop an efficient query answering algorithm via index traversal. We verify the efficiency and effectiveness of our pruning techniques, indexing mechanism, and query processing algorithm through extensive experiments on real-world and synthetic data sets under various parameter settings.