Adaptive Preference Intelligence for Personalized Tourism Recommendation through Hybrid Computational Reasoning

Abstract

The increasing availability of digital tourism information has transformed the way travelers plan their journeys, creating a demand for intelligent systems capable of delivering recommendations tailored to individual preferences. Conventional trip planning methods and rule-based recommendation platforms generally provide the same suggestions to a broad range of users, often overlooking personal interests, financial limitations, and available travel time. This research introduces an intelligent tourism recommendation framework designed to generate customised travel options by analysing destination-related information collected from Indonesian tourism resources. To evaluate predictive capability, several machine learning algorithms with Classification and Regression Trees (CART), including Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF), are implemented as benchmark models. The proposed framework incorporates a hybrid architecture named Neuro-Tree-Net (NTN), which combines the feature-learning ability of an Artificial Neural Network with the ensemble decision-making strength of the Extra Trees Classifier (ETC) algorithm. This integration enhances the system's capacity to identify intricate patterns and hidden relationships within tourism datasets, leading to more precise recommendation outcomes. Travelers can specify criteria such as budget, trip duration, and desired destination ratings, allowing the system to generate recommendations that closely match their requirements. Data preprocessing, feature engineering, model development, and performance evaluation are carried out using Python-based libraries such as pandas, scikit-learn, TensorFlow/Keras, and Matplotlib, while an interactive Django-based web application provides seamless access to the recommendation engine. Experimental findings indicate that the proposed hybrid framework consistently achieves superior predictive performance compared with conventional machine learning techniques, offering an efficient and practical solution for personalised travel planning and tourism decision support.