作者: Song Xuekun , Zhang Han , Li Yaoting , Huo Yahui , Xiao Shaochong
DOI: 10.1109/CHICC.2015.7260995
关键词:
摘要: The identification and classification of different cancer type feature gene subset selection are great importance in diagnosis have recently received a deal attention the field bioinformatics. On basis comparing with normal samples by SVM verifying disease group can be classified vectors, we selected module types training set improved Relief algorithm, then put to test including 4 kinds samples. results series experiments conditions proved that accuracy genes reach more than 95%. algorithm show excellent performance selecting identify classify types.