Defining and detecting k-bridges in a social network: The Yelp case, and more

作者:

Highlights:

摘要

In this paper, we introduce the concept of k-bridge (i.e., a user who connects k sub-networks of the same network or k networks of a multi-network scenario) and propose an algorithm for extracting k-bridges from a social network. Then, we analyze the specialization of this concept and algorithm in Yelp and we extract several knowledge patterns about Yelp k-bridges. In particular, we investigate how some basic characteristics of Yelp k-bridges vary against k (i.e., against the number of macro-categories which the businesses reviewed by them belong to). Then, we verify if there exists an influence exerted by k-bridges on their friends and/or on their co-reviewers. We also analyze the relationship between k-bridges and power users. In addition, we investigate the relationship between k-bridges and the main centrality measures in the macro-categories of Yelp. We also propose two further specializations of k-bridges, regarding Reddit and the network of patent inventors, to prove that the knowledge on k-bridges we initially found in Yelp is not limited to this social network. Finally, we present two use cases that can highly benefit from the knowledge on k-bridges detected through our approach.

论文关键词:k-bridge,k-bridge detection algorithm,Multi-network scenario,Influencers,Analysis of co-reviewers,Yelp,Reddit,PATSTAT-ICRIOS

论文评审过程:Received 22 October 2019, Revised 5 February 2020, Accepted 1 March 2020, Available online 4 March 2020, Version of Record 4 April 2020.

论文官网地址:https://doi.org/10.1016/j.knosys.2020.105721