Clustering Analysis of Gene Data Based on PCA and SOM Neural Networks

A new method combined PCA (Principal Component Analysis) with SOM (Self-Organizing Maps) neural network is presented for clustering analysis of gene expression data. Firstly, the principal components are extracted from the genetic data set by PCA, in order to get a low dimensional data set. These principal components with lower dimension can basically express comprehensive information of original data set. Secondly, the features from principal components are clustered by SOM, the similar gene data are grouped into same area. Compared with Self-Organizing Maps (SOM), the integrated PCA-SOM method can obtain a higher correct clustering rate and clear boundary. The experimental results show that the performance of new method for the clustering analysis of gene expression data is efficient and effective.