Modelling dense relational data

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

    Abstract

    Relational modelling classically consider sparse and discrete data. Measures of influence computed pairwise between temporal sources naturally give rise to dense continuous-valued matrices, for instance p-values from Granger causality. Due to asymmetry or lack of positive definiteness they are not naturally suited for kernel K-means. We propose a generative Bayesian model for dense matrices which generalize kernel K-means to consider off-diagonal interactions in matrices of interactions, and demonstrate its ability to detect structure on both artificial data and two real data sets.
    Original languageEnglish
    Title of host publication2012 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)
    Number of pages6
    Place of Publication978-1-4673-1025-3
    PublisherIEEE
    Publication date2012
    ISBN (Print)978-1-4673-1024-6
    DOIs
    Publication statusPublished - 2012
    Event2012 IEEE International Workshop on Machine Learning for Signal Processing (MLSP) - Santander, Spain
    Duration: 23 Oct 201226 Oct 2012
    http://mlsp2012.conwiz.dk/

    Conference

    Conference2012 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)
    CountrySpain
    CitySantander
    Period23/10/201226/10/2012
    Internet address
    SeriesMachine Learning for Signal Processing
    ISSN1551-2541

    Keywords

    • Relational Modelling
    • Non-parametrics
    • Infinite Relational Model
    • Granger Causality

    Cite this