Data-Driven Insights In To Diabetes Risk Factors And Patient Care Optimization Using Machine Learning Algorithms

Given the rising incidence of type2 diabetes worldwide, it is becoming more and more crucial to find possible origins of the condition. Finding possible reasons for this illness is still quite difficult, however. The goal of this effort is to use machine learning (ML) algorithms to predict diabetes risk factors. Diabetes mellitus is a chronic condition characterized by abnormalities in the metabolism of proteins, fats, and carbohydrates. The primary sign of diabetes is hyperglycemia or elevated blood sugar levels. This paper outlines a supervised learning technique to create efficient risk prediction tools fortheoccurrenceoftype2diabetes. A features analysis is used to examine their importance and possible connection to diabetes. The feature extraction techniques accomplished in principal component analysis (PCA) have been used to increase prediction accuracy. These elements, corresponding to the most common symptoms of diabetes that might appear gradually, are used to train and evaluate several machine learning models. Precision, Accuracy, Recall, F-Measure, and AUC are the measures used to assess and compare various machine learning models under a 5-fold cross-validation and data splitting. Both validation methods indicated that the Graph Convolutional Network (GCN) algorithms were the best-performing models.