Investigation of Machine Learning-Based Algorithms for Identification of Diabetes Disease

Diabetes, a condition characterized by disrupted insulin production or insulin resistance, results in elevated blood glucose levels. It can lead to severe issues such as heart disease and kidney failure. Early prediction of diabetes is crucial for mitigating risks and managing the severity of the condition. The paper presents effective algorithms of machine learning (ML) for diabetes prediction, including Decision Trees (DT), K-Nearest Neighbors (KNN), Light Gradient-Boosting Machine (LGBM), Support Vector Machine (SVM), and Random Forest (RF), all based on ML algorithms. Various ML models’ performance was assessed using the PID Indian Diabetes (PID) dataset, encompassing parameters like age, glucose levels, insulin levels, and skin thickness, among others. The examination unveiled that among all models, the LGBM algorithm showcased superior performance, attaining an impressive accuracy of 92%. Consequently, LGBM emerges as the most efficient algorithm for accurately distinguishing between diabetic and non-diabetic individuals.