Social Networks Discovery Based on Information Retrieval Technologies and Bees Swarm Optimization: Application to DBLP

Social Networks Discovery Based on Information Retrieval Technologies and Bees Swarm Optimization: Application to DBLP

Yassine Drias (USTHB, LRIA, Algeria) and Habiba Drias (USTHB, LRIA, Algeria)
DOI: 10.4018/978-1-4666-9562-7.ch043
OnDemand PDF Download:
List Price: $37.50
10% Discount:-$3.75


Unlike the previous works where detecting communities is performed on large graphs, our approach considers textual documents for discovering potential social networks. More precisely, the aim of this paper is to extract social communities from a collection of documents and a query specifying the domain of interest that may link the group. We propose a methodology that develops an information retrieval system capable to generate the documents that are in relationship with any topic. The authors of these documents are linked together to constitute the social community around the given thematic. The search process in the information retrieval system is designed using BSO, the bee swarm optimization method in order to optimize the retrieval time for large amount of documents. Our approach was implemented and tested on CACM and DBLP and the time of building a social network is quasi instant.
Chapter Preview

Social networks mining has been widely studied this last decade and several works have reached important advancements in the area. Applications to domains like biology, economics and marketing have been also undertaken and the results were promising. Among the main issues that interest the researchers are the mining methods, community identification and modeling social rating networks. Lots of investments such those of (Clauset et al. 2004, Flake et al. 2000, Fortunato 2010, Newman & Girvan 2004, Radicchi et al. 2004) have been devoted to community identification. All of these articles build the social network from a given large graph and differ from each other by the method designed to extract the community structure. In (Flake et al. 2000), the authors focused especially on web communities and in (Domingos & Richardson 2001), the authors considered the marketing application. Other axes that were also investigated concern the analysis of the social network as in (Leskovec et al. 2008) and the scoring and evaluation of the social community as in (Newman & Girvan 2004, Domingos & Richardson 2001). The community scoring function quantifies how ‘efficient’ is the community.

On the other hand, information retrieval has known extremely interesting developments for more than four decades. The general concepts and techniques are well described in (Christopher et al. 2008, Rijsbergen 1979, Salton 1976).

Finally BSO, the bee swarm optimization approach is introduced in (Drias et al. 2005) and one of its applications to web information retrieval is published in (Drias et al. 2010).

Complete Chapter List

Search this Book: