Implementation of Lifestyle Determination in Obesity Patients Using Random Forest Method

Abstract

In the digital era, the application of Internet of Things (IoT) technology for health monitoring has become increasingly important to support improved quality of life and more accurate health recommendations. This research focuses on the development of a health monitoring application integrated with a MySQL database and employing the Random Forest method for data analysis. The system collects various health-related parameters, including height, weight, eating frequency, calorie intake, water consumption, physical activity level, sleep patterns, and stress levels. All collected data are stored in a MySQL database and subsequently analyzed using the Random Forest algorithm to classify user lifestyles into deficient, normal, and excessive categories. The dataset is divided into training and testing data to ensure reliable model performance and accurate predictions when new data are introduced. The results indicate that the Random Forest method provides good classification performance, with an average accuracy of 68% when compared to evaluations conducted by professional nutritionists. In addition, sensor accuracy testing shows excellent measurement performance, with an average error of 0.002% for the ultrasonic sensor and 0.0204% for the load cell sensor, corresponding to accuracy levels of 99.998% and 99.9796%, respectively. Overall, the system achieves an accuracy rate of 98.65%, demonstrating that the integration of IoT technology, database systems, and machine learning methods is effective and has strong potential for broader health monitoring applications.