Prediction of obesity among school going children using Machine learning algorithms

Abstract

Obesity has developed as a major global concern in the modern era due to the continuing trend towards unhealthy lifestyles characterized by excessive junk food consumption, insufficient sleep patterns, and prolonged periods of sedentary behavior. This research aims to predict the possibility of obesity using machine learning algorithms. We collected data from a total of 250 samples, which included moms and school-aged children of varied ages, as well as those affected by obesity and those who were not. Six well-ML learning techniques were used to study this. We used supervised machine learning techniques including RF, DT, GB, adaboost, catboost, and MLP to evaluate the suggested model. The results showed that combining the AdaBoost and MLP algorithms produced the best results, with 96% accuracy, 92% precision, 96% recall, and 94% F1. Using machine learning to analyze public health data can improve forecasts, find complicated patterns, and improve our understanding of issues such as identifying obesity risk factors.