Obesity Risk Prediction Using Random Forest Based on Eating Habit Parameters

Abstract

Obesity is a global health problem associated with multiple chronic diseases, so early detection and risk prediction are important for prevention efforts. As obesity is one of the major health problems that can lead to various chronic diseases, accurate modelling can help in prevention and early intervention efforts. This study aims to develop an obesity risk prediction model using Random Forest technique, which is based on individual eating habit parameters. The dataset used is taken from an open dataset that has variables of eating habits, which includes variables such as frequency of consumption of high-calorie foods, eating patterns, and types of food. The data was processed and analysed with the Random Forest algorithm, an ensemble learning method known to be effective in handling datasets with high dimensionality and non-linear relationships between features. The developed Random Forest model showed good performance with a prediction accuracy of 81.76%. This accuracy indicates that the model can effectively distinguish individuals with high risk of obesity from those with low risk based on their eating habit parameters. The results of this study demonstrate the potential of Random Forest as a useful tool in identifying obesity risk, which can assist in data-driven health prevention and intervention strategies.