16-LQP-ARJCS 20251231000000 Lumina Quest Publishing m.arslansohail@gmail.com Lumina Quest Publishing Advanced Research Journal of Computer Science Adv. Res. J. Comput. Sci. 3134-884X 10.XXXXX/arjcs https://lquestpub.com/archives.php?journal=advanced-research-journal-of-computer-science 12 31 2025 12 31 2025 2 2 10.XXXXX/arjcs20250102 https://lquestpub.com/articles-list.php?journal=advanced-research-journal-of-computer-science&volume=2&issue=2 Predictive Modeling for Diabetes Risk Assessment Using the Healthcare Diabetes Dataset: A Machine Learning Approach Maham Nasir Safina Shahzadi Muhammad Usman Muhammad Tehseen Qureshi Diabetes mellitus remains a major global health challenge, with early prediction critical for intervention. This study leverages the Healthcare Diabetes Dataset [1] - comprising 768 records of female patients of Pima Indian heritage - to develop a binary classification model for diabetes diagnosis (Outcome: 0 = No, 1 = Yes). Key attributes include Pregnancies, Glucose, Blood Pressure, Skin Thickness, Insulin, BMI, Diabetes Pedigree Function and Age. After addressing disguised missing values (zeros in clinically impossible fields), Exploratory Data Analysis (EDA), feature scaling and an 80/20 train-test split, a Logistic Regression model was trained and evaluated. The model achieved 77.26% accuracy, with strong performance on the majority class (non-diabetic: Precision 0.79, recall 0.90) but moderate recall on the diabetic class (0.52), reflecting typical challenges with class imbalance. Results align with recent literature (70–80% range for Logistic Regression on this dataset). This paper provides a complete, reproducible pipeline, discusses ethical considerations, compares with advanced methods and suggests improvements. It serves as an educational benchmark for healthcare predictive analytics. 12 31 2025 12 31 2025 35 38 10.XXXXX/arjcs2025010201 https://lquestpub.com/article/10.XXXXX/arjcs2025010201