| MS Seminar

Name of the Speaker: Ms. C K Vaishnavi (EE23S028)
Guide: Dr. Mohanasankar S
Online meeting link: https://meet.google.com/gha-xdte-sfw?hs=224
Date/Time: 3rd April 2025 (Thursday), 2:15 PM
Title: A Framework for VO2max Prediction in Gamified Cardiac Assessment using Feature-Driven Machine Learning Models.

Abstract :

Maximal oxygen uptake, VO2max, is a key measure of cardiovascular fitness, traditionally assessed via resource-intensive Cardiopulmonary Exercise Testing (CPET). To provide a scalable alternative, this study introduces the Cardiopulmonary Spot Jog Test (CPSJT)—a gamified, mobile-based assessment using a chest-worn heart rate and motion sensor. Feature selection, based on statistical correlation analysis and model performance, identified a minimal yet effective predictor set. Machine learning models trained using these features, validated through Stratified 5-Fold Cross-Validation, demonstrated strong predictive performance, with linear regression achieving an RMSE of 5.78 and training-test correlations of 0.83 and 0.84. While Random Forest and Support Vector Regression showed improvements, all models exhibited minimal overfitting. These findings highlight CPSJT’s potential as a practical, resource-efficient alternative to CPET, enabling scalable VO2max estimation through mobile and wearable technologies.