This study investigated lifestyle behaviours among 306 medical undergraduates (153 first-year MBBS students and 153 interns) to identify early risk factors for non-communicable diseases and evaluate the role of artificial intelligence (AI) and machine learning (ML) in predicting lifestyle-related health risks. Data were collected using validated questionnaires assessing sleep quality (PSQI), physical activity (GPAQ), and dietary habits (AFHC), along with anthropometric measurements obtained through bioelectrical impedance analysis.

The findings showed that interns had higher body weight, BMI, muscle mass, visceral fat, basal metabolic rate, healthier dietary scores, and greater physical activity than first-year students, but also significantly poorer sleep quality. Male participants demonstrated higher weight, muscle mass, visceral fat, and BMR, while females had higher body fat percentage. Machine learning models, particularly Random Forest (approximately 88% accuracy), outperformed conventional approaches in predicting lifestyle-related risk patterns by identifying complex interactions among diet, sleep, physical activity, and body composition. The study highlights the value of integrating AI with traditional epidemiological methods and emphasizes the need for phase-specific lifestyle interventions within medical education to reduce future cardiometabolic risk.