A hybrid Cancer Classification Based on SVM Optimized by PSO and Reverse Firefly Algorithm
Abstract
Feature selection is used by all researchers to identify the features of genes from the high dimensional dataset. It provides a pathway to identify the actual phase of the disease so that the accurate precaution can be considered to save a life. In this article, to identify the irrelevant as well as the redundant features we have adopted FCBF (First Correlation-based feature selection). Later we have considered SVM which is optimized by PSO and recursive FA(Firefly algorithm). Here the adopted method is known as PRFA-SVM. The proposed approach is applied to different well-known datasets publicly available in repositories. The comparison of classification accuracy states that the proposed PRFA-SVM approach provides effectiveness and robustness.