A novel SVM-RFE based biomedical data processing approach: Basic and beyond

Abstract

Standard support vector machine (SVM) serves as a powerful method for faults diagnosis and classification for last two decades. However, this standard method and its recent versions still encounter many problems for classification and diagnosis. To cope with these problems, a method (SVM-RFE) combines SVM, recursive feature elimination (RFE) and principal component analysis (PCA) is presented in this paper. SVM-RFE and PCA are able to effectively reduce the dimension of features and extract the most relevant features. Based on them, the classification accuracy and calculation speed are improved. In order to demonstrate the effectiveness of the proposed methods, they are applied to diagnose and classify the Wisconsin Diagnostic Breast Cancer (WDBC) data sets with 30 features. It should be noticed that SVM-RFE and PCA are initially applied to this data and the classification accuracy is much better than many other researchers' results. It can be seen that the proposed methods show satisfactory simulation results not only for diagnosing breast cancer but also for its high classification accuracy.