Title: Functional classification for high dimensional data
Authors: Yuko Araki - Shizuoka University (Japan) [presenting]
Abstract: Classification of high dimensional data, such as images or longitudinal data, is considered by using basis expansions with the help of subspace methods. The proposed method is capable of classifying very high dimensional data by using composite basis expansions with sparse principal component analysis. The selection of an appropriate dimensions of decision space are selected by using model selection criteria. The proposed methods are applied to real image data analysis and Monte Carlo simulations are conducted to examine the efficiency of our modelling strategies comparing to other classification methods.