Diabetes Prediction Using Gaussian Naive Bayes and Artificial Neural Network

Early detection of diabetes can help prevent possible complications such as heart problems, kidney problems, nerve damage, and vision loss. Early detection not only improves health status but also improves overall quality of life by allowing timely control of blood sugar levels and important lifestyle changes. The main goal is to develop improved cohort-based diabetes prediction models. The main goal is to allow early diagnosis and early initiation of treatment. The strategy incorporates key steps such as comprehensive data preprocessing, exploratory data examination, and application of dimensionality reducing methods such as vital component investigation (PCA). This involves building predictive models using both Artificial Neural Networks (ANN) and Gaussian Naive Bayes (GNB). Additionally, a cluster model has been developed that allows comparative analysis between the different techniques used. The proposed model has a promising accuracy of approximately 94% for predicting diabetes onset, which has attracted attention and high-lights its possible importance in real-world errors. Emphasizing detailed data analysis, data preprocessing optimization, rigorous model training, and comprehensive performance analysis. These learning efforts, combined with sophisticated data and advanced analytical concepts, deliver the results you expect from well-trained individual and cluster models.