Population-Based Feature Selection for Biomedical Data Classification

Abstract

Classification of biomedical data plays a significant role in prediction and diagnosis of disease. The existence of redundant and irrelevant features is one of the major problems in biomedical data classification. Excluding these features can improve the performance of classification algorithm. Feature selection is the problem of selecting a subset of features without reducing the accuracy of the original set of features. These algorithms are divided into three categories: wrapper, filter, and embedded methods. Wrapper methods use the learning algorithm for selection of features while filter methods use statistical characteristics of data. In the embedded methods, feature selection process combines with the learning process. Population-based metaheuristics can be applied for wrapper feature selection. In these algorithms, a population of candidate solutions is created. Then, they try to improve the objective function using some operators. This chapter presents the application of population-based feature selection to deal with issues of high dimensionality in the biomedical data classification. The result shows that population-based feature selection has presented acceptable performance in biomedical data classification.