作者: Xiaohui Lin , Jiuchong Gao , Lina Zhou , Peiyuan Yin , Guowang Xu
DOI: 10.1016/J.JCHROMB.2014.05.044
关键词:
摘要: In systems biology, the ability to discern meaningful information that reflects nature of related problems from large amounts data has become a key issue. The classification method using top scoring pairs (TSP), which measures features set in and selects ranked feature construct classifier, been powerful tool genomics analysis because its simplicity interpretability. This study examined relationship between two features, modified ranking criteria k-TSP measure discriminative each pair more accurately, correspondingly, provided an improved procedure. Tests on eight public sets showed validity method. was applied our serum metabolomics derived liquid chromatography-mass spectrometry hepatocellular carcinoma chronic liver diseases. Based 27 selected pairs, HCC diseases were accurately distinguished principal component analysis, certain profound metabolic disturbances disease development revealed by pairs.