Functional discriminant analysis for gene expression data via radial basis expansion

機関リポジトリ (HANDLE) オープンアクセス

説明

In this paper we introduce functional discriminant analysis which is an extension of the classical method of logistic discriminant analysis to the data where predictor variables are functions or curves. The functional discriminant analysis approach can classify curves belong to two distinct classes effectively by imposing smoothness constraint on the predictor functions and coefficient function via regularized radial basis expansion. In order to select the number of basis functions to be expanded and the value of smoothing parameter which are essential in regularization, we derive an information criterion which enables us to evaluate model estimated by regularization. The proposed method is illustrated with the example in the analysis of yeast cell cycle microarray data. It is shown that functional discriminant analysis performs well especially in the sense of prediction accuracy.

収録刊行物

詳細情報 詳細情報について

  • CRID
    1050298532705849472
  • NII論文ID
    120006654336
  • HANDLE
    2324/11831
  • 本文言語コード
    en
  • 資料種別
    conference paper
  • データソース種別
    • IRDB
    • CiNii Articles

問題の指摘

ページトップへ