Abstract: Cluster analysis is a popular method for gene expression data,which can be used for finding cancer cell so that the diseases can be diagnosed accurately and rapidly through the gene class label.However,more attributes and less samples will produce a mass of redundant or disturbed information,resulting in the decline of the accuracy of the direct clustering in high dimensional data.Principal Component Analysis(PCA) is a classical method for dimension reduction which can transform high dimension data into low space under maintaining maximal variance.The shortcoming of PCA is the lack of strong interpretation for loadings that have no characteristic of sparsity.In this paper,a sparse PCA methodbased on Truncated Power was applied into the feature extraction for gene expression data,then the sparse PCA was fed into K-means process for clustering.Finally,the experimental results on Colon cancer,leukemia and lurg cancer three typical gene datasets verify that the sparse gene data can improve the efficiency and accuracy on clustering.

Key words: Gene expression data,Loadings,Truncated power,Sparse principal component analysis,Feature extraction