Analysis and Evaluation of TripAdvisor Data: A Case of Pokhara, Nepal

Tan, Wenan; Shrestha, Deepanjal; Gaudel, Bijay; Rajkarnikar, Neesha; Jeong, Seung Ryul

doi:10.1007/978-3-030-93247-3_71

Cited by 5 publications

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Social media platforms have played an indispensable role for people to discuss and convey their experiences and opinions regarding various aspects of life. Such platforms cater to a massive amount of data useful for the travel and tourism industry [2], [3], [8], [14]. The United Nations has recognized tourism as a major economic activity that can make a significant contribution to achieving Sustainable Development Goals (SDGs) [15].…”

Section: Introductionmentioning

confidence: 99%

Machine Learning-Based Social Media Review Analysis for Recommending Tourist Spots

Lahagun,

Devkota,

Giri

et al. 2024

J. Eng. Sci.

View full text Add to dashboard Cite

In recent years, the tourism industry has witnessed significant growth, resulting in an increased demand for effective and personalized tourist place recommendation systems. In this study, a tourist spot recommendation system is proposed which is built by developing a machine learning model based on a Support Vector Machine (SVM), Decision Tree (DT), and k-Nearest Neighbors(k-NN). Public experiences and opinions regarding the various spots available in popular social media sites such as TripAdvisor, Google, Instagram, and TikTok are utilized to train the model. The system matches the probability of the user query with the predicted probability of reviews for a particular spot. The SVM algorithm, known for its robustness in handling high-dimensional data, is adapted to model the complex relationships between users' reviews, spots, and their attributes. Real-world data is used to evaluate the system's performance, demonstrating its ability to significantly improve the user experience and contribute to the sustainable growth of the tourism sector. The system's capability was demonstrated as it achieved a notable F1-Score of 0.78 when SVM was implemented. Additionally, a promising accuracy rate of 93.023% was observed when random queries were used for tourism spot prediction, emphasizing that SVM outperformed DT and k-NN.

show abstract