Comparative analysis for calorie burn prediction using machine learning techniques

Predicting calorie burn is emerging as a crucial field of research in the development of fitness and health technologies. This study uses sophisticated machine learning (ML) algorithms-Five regression models, including CATBoost, XGBoost, Random Forest (RF), KNN, and SVM, have been utilized to discover the optimal approach. In this CATBoost regression model performed exceptionally well, with the lowest Mean Absolute Error (MAE) of 0.48, Mean Squared Error (MSE) of 0.469, R² Score of 0.9998, Mean Absolute Percentage Error (MAPE) of 0.012, Mean Bias Error (MBE) of 0.030, Median Absolute Error (MAR) of 0.38, Relative Absolute Error (RAE) of 0.008. This research investigates the use of machine learning to estimate calorie burn. These evaluation metrics provide a comprehensive analysis of the models‘ accuracy and reliability in estimating calorie expenditure. The proposed methodology involved the computation of evaluation metrics such as Mean Bias Error, Median Absolute Error, and Relative Absolute Error, with the CATBOOST regression model demonstrating the most favorable performance. These findings hold significant potential for advancing wearable technologies, enabling the creation of personalized fitness regimens, and contributing to the prevention of overweight and obesity through enhanced insights into calorie expenditure.