Implementing Calorie Burnt Prediction Through XGBOOST Algorithm Compared Over Random Forest Regressor

This research aims to evaluate the XGBoost algorithm and the Lasso Regression approach for their respective abilities to forecast caloric expenditure during physical activity. Tools and Techniques for Scientific Investigation: Information such as calorie burn rates from different activities are included into the planned research. The XGBoost technique and state-of-the-art Lasso Regression were among the machine learning technologies used to ascertain these calories. The XGBoost method employs Innovative Lasso Regression sampling with a 10-sample size and a 0.8 G power value. According to an independent sample T-test, the two methods’ accuracy varies by 0.001 ($\mathbf{p}{\lt}0.05$). The average of the two approaches is 72.26, and the average detection accuracy is $\pm 2 S D$. In conclusion, the innovative Lasso Regression approach achieved a 72.35% accuracy rate, but the XGBoost strategy achieved an accuracy rate of 83.73%.