Multi-class logistic discrimination via wavelet-based functionalization and model selection criteria

IR (HANDLE) Open Access

Description

We consider a multi-class logistic discrimination for functional data. We use a wavelet-based smoothing technique in obtaining a set of functional data, from irregularly sampled time-dependent covariates of a number of individuals. A method of estimating discriminant model is based on a regularized log-likelihood, where we apply model selection criteria derived from Kullback-Leibler information and Bayes’ analysis.

Journal

  • MHF Preprint Series

    MHF Preprint Series 2006-25 2006-07-27

    Faculty of Mathematics, Kyushu University

Keywords

Details 詳細情報について

  • CRID
    1050861482657649536
  • HANDLE
    2324/3393
  • Text Lang
    en
  • Article Type
    journal article
  • Data Source
    • IRDB

Report a problem

Back to top