Aug 2026· Journal of Organizational and End User Computing· Vol 38, pp. 1-25· 0 citations
TL;DR
An integrated framework combining Topic-RoBERTa, topic-level semantic network analysis, and Graph Attention Networks (GAT) to extract experience topics and model their semantic associations and offer valuable insights for tourism management and service optimization is presented.
Abstract
Online tourist reviews, a major form of user-generated content (UGC), are often short and unstructured, complicating the identification of tourist experience dimensions and their relationships. This study presents an integrated framework combining Topic-RoBERTa, topic-level semantic network analysis, and Graph Attention Networks (GAT) to extract experience topics and model their semantic associations. The authors validated the framework on 186,429 reviews from the OD-TripM TripAdvisor review dataset released by the Data Science and Computational Intelligence (DaSCI) research group on GitHub and the Yelp Open Dataset, identifying 12 tourist experience dimensions and constructing a community-structured semantic network. The results show that the proposed method can extract interpretable topics, reveal fine-grained attention-weighted associations, and provide a more structured understanding of tourist experiences. The findings offer valuable insights for tourism management and service optimization.
The rapid growth of digital tourism platforms has transformed how tourists search for travel information and select destinations. However, the abundance of available information often causes information overload, making it difficult for users to identify destinations that match their preferences. This study designs and implements a web-based tourist destination recommender system using a Hybrid Recommender System that combines Content-Based Filtering (CBF) and Collaborative Filtering (CF) to generate accurate and personalized recommendations. The CBF component applies TF-IDF and Cosine Similarity to analyze destination attributes, while the CF component predicts user preferences based on historical rating patterns. The recommendation scores from both methods are integrated using the Weighted Hybrid approach with a weighting parameter of α = 0.5. System performance was evaluated using the Precision@3 metric on a dataset containing 50 tourist destinations in Blora Regency, 50 users, and five tourism categories. The experimental results achieved an average Precision@3 score of 63.33%, indicating that approximately two of the top three recommended destinations were relevant to user preferences. These findings demonstrate that combining destination content with user rating behavior enhances recommendation quality compared with using a single recommendation technique. The proposed system can support tourists in making informed travel decisions through personalized and relevant destination recommendations while providing a practical approach for tourism recommendation applications.
Tourism has grown rapidly. Travelers face abundant attractions and heterogeneous online content. This increases the cognitive effort of trip planning and motivates personalized recommendation. Existing recommender systems emphasize objective features but they exploit subjective review signals superficially, overlooking user personality cues and attraction-level signed sentiment. They also tend to treat users’ visits as uniformly positive feedback, even though negative preferences are common and informative in textual reviews. To address these issues, we propose VEGAR, a tourist attraction recommender system that jointly models signed feedback within a signed spatial-sentiment knowledge graph. To exploit subjective signals from reviews, VEGAR introduces a subjective feature extraction module and designs a signed spatial-sentiment knowledge graph construction framework. It extracts personality-related user features and attraction signed sentiment features from textual semantics. These subjective signals are organized together with functional features and spatial features in three subgraphs: a user subgraph, an attraction positive subgraph, and an attraction negative subgraph. To model signed feedback and align it with signed sentiment, VEGAR proposes a prediction module with five propagation procedures over three subgraphs to learn signed user and attraction preference profiles. VEGAR further introduces a high-order cross aggregation module and gated fusion to achieve informative user and attraction embeddings. To capture spatial constraints while preserving semantic relevance, VEGAR proposes spatial-sentiment attention that jointly models semantic relevance and distance-based influence during propagation. Experiments on three tourism datasets, namely the Suzhou, Beijing, and Alaska datasets, show that VEGAR achieves the best performance in both CTR prediction and Top-K recommendation. Additional ablation, robustness, cold-start, and interpretability analyses further validate the effectiveness and generalizability of the proposed framework.
Renjun Cao, Yong Gao, Yi Zhang et al.· Journal of King Saud Univers...· 0 citations
User-generated content has become an increasingly relevant source of evidence for understanding visitor experiences in urban destinations. Although online reviews are widely used by tourists when making decisions, their potential as a strategic tool for destination and attraction management remains underexplored, particularly in heritage-rich cities facing growing tourism pressure. This paper analyses TripAdvisor reviews of three emblematic attractions in Porto, Portugal: Livraria Lello, Mercado do Bolhão and Torre dos Clérigos. The study is based on 150 reviews, with 50 reviews selected for each attraction, and combines descriptive quantitative analysis with qualitative thematic coding. The results show that each attraction presents a distinct experiential profile. Livraria Lello is marked by a strong tension between architectural admiration and dissatisfaction related to overcrowding, queuing, prices and perceived loss of authenticity. Mercado do Bolhão receives predominantly positive evaluations associated with food, atmosphere, renovation quality and sensory appeal, although some comments express concerns about touristification. Torre dos Clérigos is valued mainly for its panoramic views and symbolic role in the city, while the physical effort of the climb is both a constraint and part of the experience. The paper argues that user-generated content can function as a strategic diagnostic tool, helping managers identify gaps between projected destination images and lived visitor experiences. The findings provide practical insights for visitor flow management, perceived value and the preservation of accessibility and authenticity in urban tourism attractions.
Jacqueline Galdón Jiménez, Helena Albuquerque, Fátima Matos Silva· New Trends in Sustainable Bu...· 0 citations
Cruise tourism offers travelers the opportunity to visit multiple destinations while enjoying various onboard services. It has emerged as one of the fastest growing sectors in the tourism industry. Online reviews have become an essential resource for both travelers and cruise operators, with the rapid increase in user-generated content on online platforms. This study presents a novel framework for analyzing cruise passenger satisfaction across European itineraries by integrating transitive clustering, topic modeling, and Aspect Based Sentiment Analysis (ABSA). Cruises were grouped into four major geographic regions based on the ports they visited. Eight key aspects from online reviews were identified where three of them are itinerary related. The GPT-4o mini large language model, accessed through the OpenAI API, was used for fine-grained sentence-level aspect-based sentiment analysis. Results show that cruises in the Eastern Mediterranean received the highest satisfaction scores across itinerary related aspects while those in the Western European and Atlantic Coastal scored lowest. This study highlights the effectiveness of the GPT-4o mini large language model in extracting nuanced sentiment from informal online reviews. It also identifies the most and least satisfying cruise regions in Europe based on customer feedback, while providing an aspect×region knowledge map for decision support.
Ahmet Cumhur Öztürk, Gözde Öztürk· Kahramanmaraş Sütçü İmam Üni...· 0 citations
Personalisation of product rankings in e-commerce is needed because different users have different interests, demands and browsing conditions. A user-item network model can be employed to represent clicks, favourites, additions to shopping carts, ratings and purchases for personalised Top-k ranking in this paper. This review connects PageRank and Personalised PageRank with graph collaborative filtering, multi-behaviour graph learning, temporal self-supervision and industrial pre-ranking systems. According to the above research, different types of behaviour, the order of interaction and time information, graph construction methods, negative sampling, etc., can affect the ranking results. However, there are still many problems such as scarce and noisy implicit feedback, cold start, popularity bias, preference drift, scalability, privacy, fairness and lack of interpretability. In the future, research will continue to be carried out in the field of combining dynamic heterogeneous graphs, multi-behavioural objectives, causal and self-supervised learning, privacy-preserving computation and graph foundation models. In short, the user-item network model applies linear algebra, probability theory, graph propagation and representation learning to solve the problem of personalised e-commerce ranking. It is convenient to conduct a comparison of the models and decide which one to use.
Nianying Li· Theoretical and Natural Scie...· 0 citations
Bandung Raya offers hundreds of tourist destinations spread across Bandung City, Bandung Regency, and West Bandung Regency, which causes information overload and makes it difficult for tourists to select destinations that match their preferences. Recommender systems, as one of the most widely applied branches of machine learning, provide a way to filter such information automatically. This study develops a machine learning-based recommender system for tourism destinations in Bandung Raya using Content-Based Filtering, in which destination descriptions are represented as numerical vectors through TF-IDF weighting and compared using Cosine Similarity. The dataset consists of 331 destinations with name, description, category, and region attributes. Text preprocessing is performed in five stages using the Sastrawi library for the Indonesian language, producing a TF-IDF matrix of 331 by 583 and a similarity matrix of 331 by 331. The model is deployed as a website using Flask as the backend, React as the frontend, and a REST API as the interface, supporting both name-based search and free-text query search. Functional validation uses Black-Box Testing, while recommendation quality is measured using Precision at K and Mean Average Precision. All eight functional scenarios passed, with Precision at 5 of 96.00 percent, Precision at 10 of 90.00 percent, and Mean Average Precision of 98.86 percent, indicating that, for the five evaluated queries, relevant destinations are ranked highly.
Mira Aldina, Galih, Siti Nur· Bulletin of Intelligent Mach...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
