Gender-Based Comparative Study of Type 2 Diabetes Risk Factors in Kolkata, India: Frequentist versus Bayesian Approach in Machine Learning

Type 2 diabetes mellitus represents a prevalent and widespread global health concern, necessitating a comprehensive assessment of its risk factors. This study aims to evaluate and compare the predictive performance of frequentist and Bayesian machine learning models in assessing the risk of Type 2 diabetes mellitus based on age, lifestyle, BMI, and waist-to-height ratio among males and females in Kolkata, West Bengal, India. The analysis utilizes data from patients observed in the outpatient consultation department of Belle Vue Clinic in Kolkata. The frequentist models employed include Random Forest (RF), and Support Vector Classifier (SVC), while their Bayesian counterparts - Bayesian Additive Regression Trees (BART), and Relevance Vector Machine (RVM) were also used. Our findings indicate that for males, BMI is the most important predictor of Type 2 Diabetes, whereas for females, Whtr is identified as the most important predictor. This study highlights gender-specific differences in risk factors for Type 2 diabetes mellitus and contributes to understanding the effectiveness of various modeling approaches in predicting risk within this population. The insights gained from this research can inform more targeted healthcare interventions and public health strategies.