Extraction of fuzzy clusters from weighted graphs

Seiji Hotta, Kohei Inoue, Kiichi Urahama

    Research output: Contribution to journalArticlepeer-review

    2 Citations (Scopus)

    Abstract

    A type of graph spectral method is proposed in which the point ensemble is divided into fuzzy clusters based on a weighted adjoint matrix. In this method, the problem of extracting a fuzzy cluster from a point ensemble is formulated as an eigenvalue problem. This is repetitively applied so that clusters are extracted successively. First, the cluster extraction method based on the similarity matrix is applied to an undirected graph, which is then expanded to a diagraph and an undirected bipartite graph, allowing the extraction of the link structure of the web page and image browsing search by keyword. Also, a type of graph drawing method is proposed in which the cluster structure is visualized by using correspondence analysis.

    Original languageEnglish
    Pages (from-to)80-88
    Number of pages9
    JournalElectronics and Communications in Japan, Part III: Fundamental Electronic Science (English translation of Denshi Tsushin Gakkai Ronbunshi)
    Volume86
    Issue number1
    DOIs
    Publication statusPublished - Jan 1 2003

    All Science Journal Classification (ASJC) codes

    • Electrical and Electronic Engineering

    Fingerprint

    Dive into the research topics of 'Extraction of fuzzy clusters from weighted graphs'. Together they form a unique fingerprint.

    Cite this