Obesity Prediction Using Machine Learning Methods: Performance Analysis of DT, RF, and ANN Models

Obesity is a condition that occurs as a result of the interaction of genetic, environmental and lifestyle factors and has become a significant health problem worldwide. Machine learning methods are effectively used to predict the risk of obesity by analyzing various parameters such as age, gender, dietary habits and physical activities of individuals. Models such as Decision Trees (DT), Random Forest (RF) and Artificial Neural Networks (ANN) are widely preferred machine learning methods for the classification and prediction of obesity. In this study, obesity classification was performed with DT, RF and ANN machine learning models using Obesity Prediction Dataset (OPD). The OPD dataset consisting of 2111 samples contains 17 features, one of which is the class feature. The data shows a relatively balanced distribution among the classes. In the study, accuracy, error matrix and processing times were analyzed to evaluate the performance of the models. According to the results obtained, the DT model reached 94.5%, the RF model 91.9% and the ANN model 91% accuracy rate. The highest classification success was achieved with the DT model, while the ANN model showed the lowest success. In addition, the ANN model had the longest test time, while the DT model was determined to be the fastest working model. These findings show that the DT model is the most successful method in the OPD data set. Artificial intelligence systems used in obesity prediction can contribute to individuals making healthier life choices by providing early diagnosis and personalized health recommendations.