PERBANDINGAN METODE ARTIFICIAL NEURAL NETWORK, DAN RANDOM FOREST PADA KLASIFIKASI TINGKAT OBESITAS

Abstract

Classification of obesity levels is an important step in supporting efforts to tackle the increasing prevalence of obesity. This study aims to compare the performance of machine learning methods, namely Artificial Neural Network (ANN) and Random Forest (RF), in classifying obesity levels based on a predetermined dataset. The research method involved data preprocessing and model training with varying proportions of training and testing data (70:30, 80:20, and 90:10). The results showed that Random Forest provided higher accuracy than Artificial Neural Network. In testing with 70% training data and 30% testing data, ANN produced an accuracy of 88.20% while RF reached 97.28%. With a training data proportion of 80% and testing data of 20%, the accuracy of ANN increased to 88.76%, while RF produced 97.37%. With a training data proportion of 90% and testing data of 10%, ANN achieved the highest accuracy of 91.39%, but it was still lower than RF, which reached 95.69%. Based on these results, it can be concluded that the Random Forest algorithm shows more optimal performance than Artificial Neural Network in obesity level classification.