14 - Multiobjective evolutionary algorithm based on decomposition for feature selection in medical diagnosis

Abstract

Disease prediction is very much important as well as crucial for the patients and also for health informatics. If the disease is detected early, then it is easy for the medical practitioners to treat early and assists them in taking more functional and preventive steps. Feature selection (FS) is being taken as a preprocessing step used to eradicate the features that are influencing negatively the efficacy of the machine learning (ML) techniques. The efficacy may be negative because of the irrelevant or may be redundant features in the dataset. Intelligent models including classification, clustering, regression, and boosting are helpful in extracting useful knowledge. Nowadays, there is an explosion in medical data, and there is also an explosion in computational technologies. Most of the medical datasets are high dimensional in nature, and so there is a need for optimal FS which is a hard problem. Hence, this chapter uses an MOEA built on decomposition (MOEA/D), which takes care of the FS in classifying medical diagnosis. This chapter presents the mathematical model of MOEA/D and its use for the FS process in medical diagnosis. The behavior of (MOEA/D) method is examined with various well-known state-of-art multiobjective optimization methods and on most of the datasets.