Departmental Bulletin Paper クラスター性によるコミュニティ抽出と情報拡散速度
Community Detection by Clustering Property and Diffusion Speed of Information
クラスターセイ ニ ヨル コミュニティ チュウシュツ ト ジョウホウ カクサン ソクド

熱田, 圭佑  ,  森口, 一郎  ,  アツタ, ケイスケ  ,  モリグチ, イチロウ  ,  Keisuke, Atsuta  ,  Ichirou, Moriguchi

19 ( 1 )  , pp.1 - 12 , 2015-09-30
To detect partial structures with high density portion of links from arbitrary networks, Matsushima proposed.method using clustering property. However, the "overlapping star-graph" problem, has been remained unsolved in this technique. In this paper we resolved it by detecting overlapping star-graph nodes firstly before dividing networks to communities. Furthermore, we investigated the effect of community structures on information diffusion. By the topology information of the detected communities, the number of simulation turns that are sufficient for information to spread over the entire networks indicated that information diffusion speeds strongly depend on the size and the properties of these communities, and that the diffusion strategy by exploiting community structures is more effective and practically executable than the random diffusion strategy.

Number of accesses :  

Other information