Computer Science and Engineering graduate; machine learning research

Performance Analysis of Supervised Learning Models for Diabetes Risk Prediction

The challenge of comparative prediction studies for diabetes must be evaluated in terms of discrimination and whether any improvement over the algorithm is maintained over time or is converted to meaningful probabilities. In this study, six supervised learning models were compared and evaluated, involving the same participants and predictors, and inference was performed using a survey in addition to a later-period test set. Methods: NHANES 2007–2018 provided 31,772 adults aged at least 20 years.

2026

A Comparative Study of Machine Learning Algorithms for Early Disease Prediction Using Public Healthcare Datasets

When predicting readmission for hospitalized people with diabetes, comparisons that maintain patient independence must be made and that separate from the discrimination of the risk estimates are their reliability and practical implications. Objective: Traditional vs. ensemble classifiers for recorded readmission within 30 days of discharge and the contribution of feature families. The UCI Diabetes 130-US Hospitals dataset was restricted to eligible discharges, which included 98,470 encounters fr

2026