<![CDATA[This study proposes an interpretable ABSA–C5.0 hybrid framework for analyzing tourists’ perceptions of digital destination information services. This framework integrates Aspect-Based Sentiment Analysis (ABSA) for aspect extraction and sentiment assessment with C5.0 decision trees for classification and rule generation. This study follows the Knowledge Discovery in Databases (KDD) process, including preprocessing, feature engineering, modeling, and evaluation. The experiment uses a statistically synthesized dataset containing 72 labeled reviews, designed to reflect real-world online review patterns. With a 70:30 validation split, the model achieved 97.22% accuracy and a Cohen’s Kappa value of 0.947 in this controlled setting. However, these results are not intended for generalization, given the limited dataset size, synthetic data construction, and the absence of baseline comparisons and statistical significance tests. The extracted rules indicate that interactivity, clarity, and response speed are the primary factors driving positive perceptions. This framework is suitable for exploratory analysis, while further validation with real-world data and comparative models is required.]]>
Copyrights © 2026
