Traditional method versus heuristic techniques for Social Network clustering

2016 Fiscal Year Final Research Report

• Project/Area Number: 26330270
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Soft computing
• Research Institution: University of Tsukuba
• Principal Investigator: MIYAMOTO Sadaaki 筑波大学, システム情報系(名誉教授), 名誉教授 (60143179)
• Research Collaborator: Torra Vicenc University of Skovde, Professor
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: ソーシャル・ネットワーク / クラスタリング / 階層的クラスタリング / メドイド法 / 2段階階層的クラスタリング / Ward法 / ファジィクラスタリング

Outline of Final Research Achievements

Analysis of social networking services (SNS) is recently focused on by many researchers and techniques of network clustering have been developed. In contrast, traditional methods of clustering are considered to be inappropriate for network clustering. In this study new methods based on traditional ideas of clustering are developed and performances are compared with known methods of network clustering. Main results of this study include the development of two-stage clustering methods using the medoids and agglomerative hierarchical clustering which is effective in SNS clustering. Moreover the Ward method of agglomerative hierarchical clustering is shown to be applicable directly to networks without the use of a positive-definite kernel.

Free Research Field

ソフトコンピューティング