Discriminant Analysis Methods for Microarray Data Classification
The studies of DNA Microarray technologies have produced high-dimensional data. In order to alleviate the “curse of dimensionality” and better analyze these data, many linear and non-linear dimension reduction methods such as PCA and LLE have been widely studied. In this paper, we report our work on microarray data classification with three latest proposed discriminant analysis methods: Locality Sensitive Discriminant Analysis (LSDA), Spectral Regression Discriminant Analysis (SRDA), and Supervised Neighborhood Preserving Embedding (S-NPE). Results of experiments on four data sets show the excellent effectiveness and efficiency of SRDA.