ML-Based Investigation of Hybrid Regressor Model for Accurate Body Fat and BMI Prediction

This paper presents a comprehensive and thorough investigation of machine learning (ML) algorithms for accurately predicting body fat and Body Mass Index (BMI). Our research, which includes a diverse set of algorithms such as Linear Regression, ElasticNet, Lasso, Ridge, Support Vector Regression (SVR), Bayesian Ridge combined with Kernel Ridge, Bayesian Ridge combined with LightGBM Regressor, Random Forest (RF) combined with Kernel Ridge, RF combined with LightGBM Regressor, and Kernel Ridge combined with LightGBM Regressor, has practical implications for the field. To demonstrate their potential in real-world applications, we rigorously evaluated their performance metrics, including R-squared, Root Mean Square Error (RMSE), recall, and F1 score. The paper's findings reveal that ensemble methods like RF Regressor and gradient boosting techniques like LightGBM Regressor outperform other algorithms in accurately predicting body fat and BMI. These models demonstrated superior predictive capabilities, exhibiting higher R-squared values, lower RMSE, and superior recall and F1 scores. This research sheds light on the effectiveness of ensemble learning in enhancing body composition analysis, providing valuable insights for healthcare practitioners and researchers seeking more precise predictive models for BMI and body fat estimation.