Leveraging Machine Learning for Classification of Obesity Risk Across Diverse Demography

Obesity is considered a global public health emergency due to its high risk of several chronic disorders, such as diabetes, hypertension, and cardiovascular disease. Due to the complexity of early risk prediction, obesity requires early interventions or classifications of individualized health management. This study demonstrates the implementation of several ML techniques to provide a comprehensive cross-frame study on the prediction of obesity risk. Application of multiple models, namely, LR (Logistic Regression), KNN (K-Nearest Neighbor), DTC (Decision Tree Classifier), GB (Gradient Boosting), MLP (Multiperceptron Network), and FNN (FeedForward Neural Network), are employed to check the performance on three benchmark datasets. The results showed that the accuracy of each model varied in the predictions, underlining both the benefits and the drawbacks of each approach in different scenarios. This study aims to develop more useful tools in clinical and preventive health, as it gives insight into comparing complex neural networks to conventional machine learning algorithms to predict the risk of obesity. The gradient boosting algorithm achieved the highest accuracy in all data sets, with a precision of 95% for Dataset 1 and 98% for both Dataset 2 and Dataset 3. This work underscores the potential of machine learning in public health and provides a foundation for policymakers and healthcare professionals to develop personalized and preventive strategies to combat obesity.