Functional Discriminant Analysis for Microarray Gene Expression Data via Radial Basis Function Networks

We introduce functional logistic discriminant analysis (FLDA) which is an extension of the classical method of logistic discriminant analysis to data where predictor variables are functions or curves. FLDA approach can effectively classify functions into two distinct classes by imposing smoothness constraint on the predictor functions and coefficient function by radial basis function expansion and regularization. In order to select the value of a smoothing parameter, we derive an information criterion which enables us to evaluate model estimated by regularization. The proposed method is illustrated through the analysis of yeast cell cycle microarray data. It is shown that FLDA performs well especially in terms of prediction ability.