A novel feature selection technique for SAGE data classification

Abstract

Computational diagnosis of cancer from gene expression data is a binary classification problem. Serial Analysis of Gene Expression (SAGE) is a sequencing technique used for measuring the expression levels of genes. Each SAGE library contains expression levels of thousands of genes (or features). It is impossible to consider all these features for classification and also the general feature selection algorithms are not efficient with this data. In this paper, a data mining technique called closed frequent itemset mining is proposed for feature selection. Subsequently these selected genes or features are used for the training and testing of two well known classifiers- Extreme Learning Machine (ELM) and Support Vector Machine (SVM). The performance evaluation of ELM and SVM classifiers shows that the proposed feature selection method works well with these classifiers. © Springer-Verlag Berlin Heidelberg 2013.