A Comparative Study of Naive Bayes and Decision Tree Models for City Recommendation Using Textual Descriptions

This study evaluates the performance of two popular machine learning models—Naive Bayes and Decision Tree classifiers—in predicting city ratings based on textual descriptions. The research utilizes a dataset containing city names, detailed descriptions, and associated ratings to assess the suitability of these cities for travelers. To process the textual data, Term Frequency-Inverse Document Frequency (TF-IDF) vectorization is applied, converting textual descriptions into numerical features for model input. Both models are trained and tested on the dataset, and their performance is compared using standard evaluation metrics, including accuracy, precision, recall, and F1-score. The findings reveal that Naive Bayes performs well in scenarios with large datasets and imbalanced data, demonstrating its computational efficiency. In contrast, the Decision Tree model offers superior interpretability, with clear decision-making paths that identify key features influencing the prediction of city ratings. This comparative analysis highlights the strengths and limitations of each model, providing insights into their applicability for personalized city recommendation systems in the tourism industry. Keywords—: Naive Bayes, Decision Tree, City Recommendation, Textual Descriptions, TF-IDF Vectorization