A survey on Diabetes Prediction Models Using Data Mining Techniques: issues and challenges.

Abstract

Diabetes is a chronic disease that affects a significant number of individuals worldwide, and timely detection and management can prevent or delay the development of severe complications. To aid in early diagnosis and treatment, data mining techniques have been extensively utilized to create predictive models for diabetes. This review paper provides an overview of recent studies on diabetes prediction models developed using data mining techniques. The review paper discusses various data mining techniques employed for diabetes prediction, such as decision trees, neural networks, logistic regression, support vector machines, and ensemble methods which combine multiple models to improve performance, have also been utilized. The paper analyzes the strengths and limitations of these techniques. The review emphasizes the significance of feature selection in enhancing the performance of diabetes prediction models. Feature selection can reduce data dimensionality, eliminate irrelevant or redundant features, and improve model interpretability. Finally, the paper presents potential areas for future research in this field, including developing more interpretable models, exploring the use of deep learning techniques, and integrating multiple data sources to enhance prediction accuracy.