PREEMPTIVE DIAGNOSIS OF DIABETES USING COMPUTATIONAL INTELLIGENCE

Globally, diabetes has been adjudged a long-term metabolic condition, with major health consequences. Finding it early is very important for lowering problems and helping patients. This research looks at how to use computer-based methods to diagnose diabetes early, using information from 93 participants . The key features analyzed include Age, Gender, Increased Weight, Urine Level, e.t.c The dataset underwent thorough preprocessing, including stratified splitting to preserve class distribution, oversampling to address minority class imbalance, and hyperparameter optimization to enhance classifier performance. Six machine learning classifiers: Random Forest, Support Vector Machine (SVM), Decision Tree, K-Nearest Neighbors (KNN), Logistic Regression, and Naïve Bayes were trained and evaluated using accuracy, precision, recall, and F1-score. Analysis revealed that Urine Level and Increased Weight were the most significant predictors of diabetes, while Age contributed moderately and Gender had minimal impact. Among the classifiers, Naive Bayes achieved the highest accuracy giving (92%) followed by KNN, SVM and logistic regression (85%) and random forest and decision Decision Tree have the lowest accuracy (81%). Stratified splitting, Oversampling and hyper-parameter tuning improved model generalization and reduced overfitting, demonstrating the effectiveness of computational intelligence techniques in handling small, imbalanced datasets. The study concludes that machine learning provides a reliable and practical approach for early detection of diabetes, enabling healthcare providers to identify high-risk individuals and implement preventive measures. The findings also underscore the critical role of Urine Level and Increased Weight as primary indicators for preemptive diagnosis. This project contributes to knowledge by establishing a 5 systematic framework for preemptive diabetes prediction, highlighting the effectiveness of classifiers, and providing insights for future studies in predictive healthcare and preventive medicine.