Additive Fuzzy Spectral Method for Cluster Analysis of Community Structure

Activity: Other › Types of Public engagement and outreach - Public lecture/debate/seminar

Description

A spectral method for fuzzy clustering will be presented. The method, ADDItive Fuzzy Spectral clustering (ADDI-FS), operates on an extension of the spectral decomposition of a square matrix and extracts clusters one by one, which allows drawing a number of model-based stopping conditions. ADDI-FS applies to several relational data types differently normalized: community network structure data, affinity between multidimensional vectors, and more conventional relational data. Particular attention will be given to the application of ADDI-FS to real-world data of similarity between research topics according to work on research centers, to build fuzzy thematic clusters based on the ACM-CCS taxonomy.
Period9 Mar 2010
Held atUniversity of London, United Kingdom