Loading...

Local Community Detection in Social

Rajabi, Arezoo | 2013

718 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 44791 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Rabiee, Hamid Reza
  7. Abstract:
  8. The fast growth of social networks and their wide range of applications have made the anal-ysis of them an interesting field of research. The growth of concern in modeling large social networksand investigation of their structural features leads studies towards community detec-tion in such networks. In recent years, a great amount of effort has been done for introducing community detection algorithms, many of which are based on optimization of a global cri-terion which needs network’s topology. However, because of big size of most of the social networks , accessing their global information tends to be impossible. Hence, local commu-nity detection algorithms have been introduced. In this approach, the community of a given node is found by using its local information. These algorithms have two primary steps and an optional step:(1) selecting the most significant node for addition to the found community,(2) decision step and (3) filtering step. A well-known problem of local community detection algorithms is entrapment in a local optimum. As a result of entrapment in a local optimum, previous proposed algorithms can’t explore all nodes in the goal community leading to com-munities with a size much smaller than real one. In this work, we propose a novel algorithm called RW. This algorithm uses a lazy random walk in node selection step. Also our proposed algorithms can escape from local minimums (maximums) by joining a group of vertices together, instead of choosing a single vertex in each iteration. Moreover, it improves found communities’ features. However, this approach increases the probability of joining neighbor communities. In order to overcome this prob-lem and find the initial node’s community, we introduce a hybrid method that uses a global algorithm in its filtering step. This algorithm that is called RW-CNM improves the RW’sprecison considerably while its recall reduces slightly.In order to evalute our proposed algorithms, we applied these methods on both of real and artifitial networks. Results shows that the found groups by proposed methods have better community structure compared to well-known and latest proposed local algorithms
  9. Keywords:
  10. Social Networks ; Community Detection ; Random Walk

 Digital Object List