Development of fuzzy rough set based feature selection algorithms for cancer data classification

Abstract

Cancer data classification is one among the most essential emerging newlineapplications in biomedical field, which helps to determine the prognosis of newlinemalignant cells. The thesis analyzes the existing methods and identifies the newlineproblem areas. It illustrates how best the optimal features and instances for newlinecancer classification could be manoeuvred with improvised techniques. It is newlinefound that there is a dearth for computation process, especially when variety newlineof datasets are being used; the feature selection process does include all newlinerelevant features to cancer classification; the search algorithms are applied to newlineselect the features to deal with global search capabilities. It is further found newlinethat such capabilities have not addressed the local search abilities and that the newlinecontinuous optimization takes place randomly. newline