Dimensionality Reduction for Prostate Cancer

Oncogene is a kind of inherent genes exists in humans´ cells. It has been recognized as a genetic disease, if the cells activated, it can make a person carcinogenesis. So, the research of digging out the useful information from gene chip is very hot in modern society. The sample size is small, high dimension, nonlinear which causes the ´dimension disaster´, so dimensionality reduction becomes the key point of prostate tumors´ classification. This paper uses Sparse principle component analysis (SPCA), Laplacian Eigenmaps and Generalized Discriminant Analysis (GDA) to classify the prostate tumors, then Support Vector Machine(SVM) is used to classify the data. Due to the experiment data, GDA gets the best result.