Review of Metaheuristic Algorithms in Feature Selection based on Parkinson Disease

A well-known method for lowering the dimensionality of datasets with unnecessary and redundant features is feature selection. Over the years, a variety of techniques and algorithms have been proposed for feature selection. Metaheuristics are an approach that could be used to search for more significant features in datasets. The purpose of this article is to evaluate the use of metaheuristic optimization algorithms as a search method in feature selection for identifying Parkinson disease patients using machine learning algorithms. Furthermore, the intention is to gain an overview of the metaheuristics and machine learning trends in this area. For this analysis, thirty-four final papers were selected. The intention of this article is to explore, evaluate, and then present a list of metaheuristics, supervised learning algorithms, performance metrics, resampling methods, statistics tests, and fitness function forms linked to PD data. Findings about each component are summarized in some final tables, and several conclusions are given.