Kernel Fisher Discriminant Analysis Embedded with Feature Selection

Abstract

As one of the state-of-the-art classification methods, kernel Fisher discriminant analysis has both theoretical advantages and successful applications. The paper proposes a new kernel Fisher discriminant analysis embedded with feature selection, which can solve both classification and feature selection in only one step. Six real-world data sets have been used to test the performance of the new embedded methods. The experimental results clearly show that the new methods can greatly reduce the dimensions of the inputs, without harm to the classification results.