Empirical Machine Learning Algorithm for Diabetic Prediction

The progress in technology within the contemporary healthcare industry has led to numerous advancements in predicting diseases. Diabetes mellitus is a condition that is on the rise across various age groups. With elevated blood glucose levels as its cause, diabetes now benefits from cutting-edge devices designed to detect it through blood samples. Diagnosing diabetes improperly might result in serious consequences including renal damage and heart attacks. This paper aims to enhance the diagnosis of diabetes by predicting blood glucose levels more than 2 h in advance. While various methodologies exist for classifying diabetes disease (DD), our chosen approach utilizes the machine algorithm for classification and prediction based on selected features. The use of Naïve Bayes, ID3, J48, SVM, Zero-R, and random forest is particularly advantageous due to its accuracy and cost effectiveness, aligning with the goal of providing a more accessible diagnosis of DD. In our proposed method, wrapper method is employed for feature extraction. Key factors including accuracy, precision, recall, and F1-score are the subject of experimental studies. The outcomes show that implementing our suggested system for diabetes diagnosis in particular is feasible.