Tourist Satisfaction Analysis of East Aceh Coastal Ecotourism Using Stacking Machine Learning and GUI Integration

Tourist satisfaction is a key indicator for evaluating service quality and improving destination management in coastal ecotourism. East Aceh has several coastal tourism destinations with strong ecotourism potential, but tourist opinions are not always systematically analyzed for service improvement. This study analyzes tourist satisfaction in East Aceh coastal ecotourism using a stacking machine learning approach and integrates the best-performing model into a Graphical User Interface (GUI). The dataset consists of 1,679 tourist opinions collected from several coastal tourism objects in East Aceh and labeled into three classes: Satisfied, Neutral, and Dissatisfied. To prevent data leakage, the dataset was first divided into training and testing sets using a 60:40 split; TF-IDF fitting and SMOTE resampling were then applied only to the training data, while the testing set was kept unchanged. The evaluated base models include Naive Bayes, Random Forest, Support Vector Machine, and XGBoost, with Logistic Regression used as the stacking meta-model. The results show that the stacking model with SMOTE achieved the best performance with 87.05% accuracy, 87% precision, 87% recall, and 87% F1-score, outperforming the individual models in this experimental setting. However, the improvement over stacking without SMOTE was modest and was not statistically tested. The best-performing model was implemented in a GUI-based prototype that demonstrates how tourist opinions can be uploaded and classified into satisfaction categories. The prototype provides a preliminary interface for presenting classification results and has not yet been validated as an operational decision-support system.