Medical data mining analyzes health data to improve patient care, especially in diabetes management—a chronic disorder affecting over 500 million people, risking eyes, kidneys, heart, and nerves due to poor glucose regulation. It processes datasets like Pima Indians Diabetes for early prediction, classifying Type 1 and Type 2 based on glucose levels and family history. Key steps include data collection from records, secure storage, pre-processing (imputation, outlier removal, normalization, encoding, oversampling for imbalance), and modeling with ML (Random Forest, SVM) or DL (DNN). Optimized pipelines achieve up to 97% accuracy, outperforming traditional methods in speed and precision via imputation, tuning, and ensembles. Recent innovations reduce complexity and enable scalable diagnosis on diverse datasets.
Babu et al. (Thu,) studied this question.