Calories burned during exercise
Pro Research Analysisby 
Searched over 200M research papers
Exercise Types and Calorie Burn: Running, Cycling, Swimming, and HIIT
Research consistently shows that running burns the most calories per hour, followed by activities like stationary bicycling, swimming, and high-intensity interval training (HIIT). HIIT is especially notable because it continues to burn calories for up to 24 hours after the workout ends, making it highly effective for overall calorie expenditure Valmiki2023Manjunathan2021.
Key Factors Influencing Calories Burned During Exercise
The number of calories burned during exercise depends on several personal and activity-related factors. Important variables include exercise duration, average heart rate, body temperature, height, weight, gender, age, and workout intensity. These factors directly impact how much energy the body uses during physical activity, and accurate calorie estimation must consider them for personalized results Salanke2024Challagundla2024Aziz2023.
Machine Learning Models for Predicting Calorie Burn
Recent studies have focused on using machine learning to predict calories burned during exercise. Models such as XGBoost, LightGBM, Random Forest, Linear Regression, and Neural Networks have been tested for their accuracy. Among these, XGBoost and LightGBM have shown high accuracy, with mean absolute errors as low as 1.27, and neural networks have also demonstrated strong performance in predicting calorie burn Valmiki2023Tan2024Salanke2024+5 MORE. These models can provide tailored recommendations and help users track their fitness progress more precisely.
Heart Rate and Non-Invasive Calorie Estimation
Heart rate is a commonly used indicator for estimating calories burned. Devices using heart rate sensors, such as IR sensors, can provide a non-invasive and cost-effective way to estimate calorie expenditure during exercise. However, while convenient, these methods may not always be as accurate as more advanced systems .
Wearable Technology and Advanced Calorie Tracking
Smartphones and smartwatches have made it easier to monitor exercise and estimate calories burned. However, traditional smartwatch-based methods that rely solely on heart rate can be inaccurate. Newer technologies, like the JoulesEye system, use thermal cameras to monitor respiration rate and have shown much lower error rates (mean absolute percentage error of 5.8%) compared to standard smartwatch methods (37.6%). This suggests that integrating advanced sensors can significantly improve calorie burn estimation in wearable devices .
Conclusion
Calorie burn during exercise is influenced by the type of activity, individual physiological factors, and workout intensity. Machine learning models, especially XGBoost, LightGBM, and neural networks, have greatly improved the accuracy of calorie burn predictions. While heart rate-based devices offer a simple way to estimate calories burned, newer technologies that use additional physiological signals, like respiration rate, provide even more reliable results. These advancements are making it easier for individuals to monitor and optimize their fitness routines with greater precision Valmiki2023Tan2024Salanke2024+7 MORE.
Sources and full results
Most relevant research papers on this topic
Advanced Calorie Burn Forecasting with XGBoost: A Comprehensive Predictive Analysis
The XGBoost regressor algorithm accurately predicts calorie burn during exercise using physiological and environmental parameters, making it a valuable tool for fitness tracking and health management.
A Multi-Model Machine Learning Approach for Monitoring Calories Being Burnt During Workouts Using Smart Calorie Tracer
A multi-model machine learning approach, particularly the neural network model, can accurately predict calorie burn during exercise, contributing to user-friendly fitness applications.
Calories Burnt Prediction Using Machine Learning Approach
The XGBboost model accurately predicts calorie burn during physical activity with 99.67% accuracy and a mean absolute error of 1.48%, potentially benefiting personalized health coaching and wellness tracking.
Feature Selection Intent Machine Learning based Conjecturing Workout Burnt Calories
The Decision Tree and Gradient Boosting regressors effectively predict workout burnt calories, with 99% retention before and after feature scaling for Anova, Correlated Feature, Variance Based, and KBest Feature methods.
DOI