Extraction of fuzzy clusters from weighted graphs

Seiji Hotta, Kohei Inoue, Kiichi Urahama

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    2 Citations (Scopus)


    A spectral graph method is presented for partitioning of nodes in a graph into fuzzy clusters on the basis of weighted adjacency matrices. Extraction of a fuzzy cluster from a node set is formulated by an eigenvalue problem and clusters are extracted sequentially from major one to minor ones. A clustering scheme is devised at first for undirected graphs and it is next extended to directed graphs and also to undirected bipartite ones. These clustering methods are applied to analysis of a link structure in Web networks and image retrieval queried by keywords or sample images. Extracted structure of clusters is visualized by a multivariate exploration method called the correspondence analysis.

    Original languageEnglish
    Title of host publicationKnowledge Discovery and Data Mining
    Subtitle of host publicationCurrent Issues and New Applications - 4th Pacific-Asia Conference, PAKDD 2000, Proceedings
    EditorsTakao Terano, Huan Liu, Arbee L.P. Chen
    PublisherSpringer Verlag
    Number of pages12
    ISBN (Print)3540673822, 9783540673828
    Publication statusPublished - 2000
    Event4th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2000 - Kyoto, Japan
    Duration: Apr 18 2000Apr 20 2000

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349


    Other4th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2000

    All Science Journal Classification (ASJC) codes

    • Theoretical Computer Science
    • Computer Science(all)


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

    Cite this