Estimation of Obesity Levels Based on Decision Trees

In recent decades, there has been increasing concern about obesity in adolescents and adults. Obesity can cause many physical health problems and affect people’s quality of life. So people are starting to look at the factors that lead to obesity and predict the emergence of obesity. This research presents an estimation of obesity levels based on eating habits, physical condition, and other factors, using a dataset found on UCI Machine Learning Repository. This dataset contains 17 attributes and 2111 records. The labels of this dataset are classified as Insufficient Weight, Normal Weight, Overweight Level I, Overweight Level II, Obesity Type I, Obesity Type II and Obesity Type III. In this research, three major methods are chosen for prediction: Decision Trees, Logistic Regression, and K Nearest Neighbor. Finally, the result obtained by Decision Trees has the best accuracy.