A systematic literature review of diabetes prediction using metaheuristic algorithm-based feature selection: Algorithms and challenges method

Diabetes is a disruption in metabolism that leads to elevated levels of glucose in the bloodstream and causes many other problems, such as stroke, kidney failure, heart, and nerve issues that are of serious concern globally. Because many researchers have attempted to build accurate Diabetes prediction models, this field has seen significant advancements. Nevertheless, performance issues are still a substantial challenge in model building. Machine Learning techniques have shown strong performance in prediction and classification tasks. Unfortunately, they often encounter challenges due to noisy features and high feature space dimensionality, significantly affecting Diabetes prediction performance. To address the problems, we can employ metaheuristic algorithm-based feature selection. However, there has been limited research on metaheuristic algorithm-based feature selections for Diabetes prediction. Therefore, this paper presents a systematic literature review of Diabetes prediction using metaheuristic algorithm-based feature selections. The data used in this study is the last ten years of published articles from 2014 to 2024. For this extensive investigation, 50 scholarly papers were gathered and analyzed to extract meaningful information about metaheuristic algorithm-based feature selections. This paper reviews metaheuristic algorithm-based feature selection, focusing on the algorithms used and the challenges faced in diabetes prediction.