出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：Recently, many studies based on microRNAs (miRNAs) showed a new aspect of cancerclassification, and feature selection methods are used to reduce the high dimensionality ofmiRNA expression data. These methods just consider the problem of where feature to class is1:1 or n:1. But one miRNA may have influence to more than one type of cancers. However,these miRNAs are considered to be low ranked in traditional feature selection methods and theyare removed at most of time. Therefore, it is necessary to consider the problem of 1:n or m:nduring feature selection. In our wok, we considered both high and low-ranking features to coverall problems (1:1, n:1, 1:n, m:n) in cancer classification. After numerous tests, information gainand chi-squared feature selection methods were chosen to select the high and low-rankingfeatures to form the m-to-n feature subset, and LibSVM classifier was used to do the multi-classclassification. Our results demonstrate that the m-to-n features make a positive impression oflow-ranking microRNAs in cancer classification since they lead to achieve higher classificationaccuracy compared with the traditional feature selection methods.
关键词：low-ranking features; feature selection; cancer classification; microRNA
