Abstract:
This research paper presents the development and evaluation of machine learning algorithms for
accurate and efficient body fat prediction. Three algorithms, namely linear regression, decision
tree, and random forest, were employed to construct predictive models. The results showcased
remarkable accuracy levels for all three algorithms. The linear regression algorithm achieved
99.9% accuracy with a mean squared error (MSE) of 7%, the decision tree algorithm achieved
99.1% accuracy with an MSE of 38%, and the random forest algorithm achieved 98.8% accuracy
with an MSE of 56%. These outcomes underscore the effectiveness of machine learning algorithms
in forecasting body fat percentage based on the utilized dataset. The high accuracy rates indicate
successful capture of underlying patterns and relationships between predictor variables and body
fat percentage. Consequently, machine learning offers a reliable and non-invasive approach to
assess body composition. The superior performance of the linear regression algorithm can be
attributed to its ability to model linear relationships, while the decision tree algorithm, adept at
handling non-linear relationships and complex interactions, also achieved impressive accuracy.
Despite a slightly lower accuracy rate, the random forest algorithm exhibited robust performance
by combining multiple decision trees. The consistently low MSE values across all three algorithms
further demonstrate their precise estimation of body fat percentage. These results have significant
implications in healthcare, fitness, and wellness, enabling early interventions for obesity-related
health risks and facilitating personalized fitness programs. Moreover, the non-invasive nature of
the predictive models makes them practical for large-scale applications. It is important to
acknowledge limitations, such as the reliance on dataset quality and representativeness, as well as
the focus on a specific set of predictor variables. Future research could explore the inclusion of
additional features to enhance accuracy and predictive capabilities. Overall, this study successfully
developed and evaluated machine learning models for body fat prediction, offering valuable
insights into their potential for body composition analysis and laying the foundation for practical
applications in various domains.
