A New Method for Linear Dimensionality Reduction

A novel class based linear dimensionality reduction method is proposed, called Class-Wise Correlation Preserving Projection (CWCPP). In CWCPP, the relation among the original gene expression data is preserved according to a certain kind of similarity between data points, which takes special consideration of both the correlation information and the class information. Different from the traditional method, i.e., Fisher Linear Discriminant Analysis (FLD), CWCPP utilizes correlation information to guide the procedure of linear projection directions searching. Experiments on yeast gene expression data and NCI gene expression data are performed to test and evaluate the proposed algorithm.