Publication

Abstract : This is increasingly important as global obesity rates continue to rise: the causes need to be understood, and the ability to predict trends determined. In this study, we use machine learning techniques to predict obesity trends given lifestyle and biometric data from a large obesity dataset. Our methodology starts with data preprocessing that includes transforming categorical variables into numerical variables and standardizes the numerical variables into such a form so that machine learning algos can work with them properly. Supervised machine learning models, such as tree based methods, neural networks and various ensemble techniques are applied, and evaluated. For determining the most effective predictive approach, these models are evaluated by several performance metrics such as accuracy, precision, recall, and F1 score; the best model is chosen based on these metrics. In addition, the study includes explainable AI technique (SHAP—SHapley Additive exPlanations; LIME—Local Interpretable Model-Agnostic Explanations) to make the results more interpretable. The results from these methods provide insight into how different lifestyle and biometric factors contribute to the prediction of obesity; such as physical activity, diet patterns, and body mass index. The results from this work demonstrate not only how machine learning can improve prediction accuracy but also how these methods can provide actionable insights into the many contributing factors to obesity.