Comparison of Decision Tree and Random Forest Algorithm Performance for Nutrition Classification in Fast Food

Abstract

Fast food has become an essential part of the busy modern lifestyle, fast food is more popular because it makes eating easy and convenient. Today's young people are very fond of instant food. However, excessive consumption of instant food can trigger various health problems, including obsessive eating patterns. This raises the need to develop more accurate analytical methods for classifying fast food nutritional data, the purpose of classification is to obtain a decision tree model that can be used to anticipate and pay attention to how variables in the data are related to each other. In comparing the performance of the Decision Tree and Random Forest Algorithms in processing fast food nutritional data, it was found that all variables were correlated. The implementation results found that both models have extraordinary capabilities. The performance of the Decision Tree and Random Forest Algorithms on the same dataset, Random Forest outperformed Decision Tree with an accuracy value of 66.67%, while Decision Tree only achieved 55.56%, indicating that Random Forest is able to provide more accurate predictions for the test data class. In addition, the characteristics of the Random Forest group, where several decision trees are combined, provide advantages in handling data complexity and improve model generalization. These results indicate that group learning can improve the performance and reliability of predictions in building classification models, especially in the case of complex datasets.