Development of Web-based Novel Machine Learning Model Using Boosting Techniques for Early Prediction of Diabetes in Indian Adults

Diabetes is a medical condition characterized by elevated blood sugar levels in the body. It occurs when the pancreas fails to produce enough insulin or the body cannot effectively use it. Currently, the detection of diabetes relies on time-consuming tests. This paper aims to develop a web-based novel hybrid machine-learning model to improve the accuracy of diabetes prediction on a dataset collected from one of the reputed hospitals in India. In this research study, two hybrid models were developed using distinct methodologies: the first one utilized a Support Vector Machine with bootstrap bagging and Reduced Error Pruning (SVMBBREP), while the second one employed a Support Vector Machine with a genetic algorithm. Furthermore, feature selection was carried out using the Minimum Redundancy Maximum Relevance (MRMR) method. The performance of four different classifiers, namely logistic regression, Support Vector Machine, K-nearest neighbor, and Naïve Bayes, was compared, and the best classifier was chosen based on the comparison results. Subsequently, two hybrid models were created. Several metrics such as sensitivity, specificity, precision, F1 score, accuracy, and Kappa were utilized to evaluate the effectiveness of individual classifiers and hybrid models. The superior hybrid model was then integrated into a web-based system designed for diabetes prediction. The SVMBBREP approach achieved an outstanding accuracy of 99.67%, surpassing the performance of the alternative boosting model. The integration of this model into a web-based system enables individuals and healthcare professionals worldwide to assess the likelihood of developing diabetes in a 24/7 environment. By providing higher accurate predictions, this system empowers users to take proactive measures for early detection and effective management of the condition. The findings of this study offer valuable support to health care providers and individuals in their efforts to combat diabetes.