Vol.14, No.4, November 2025.                                                                                                                                                                          ISSN: 2217-8309

                                                                                                                                                                                                                        eISSN: 2217-8333

 

TEM Journal

 

TECHNOLOGY, EDUCATION, MANAGEMENT, INFORMATICS

Association for Information Communication Technology Education and Science

 

BERT Model and Word Embedding for Opinion Retrieval on Hotel Reviews

 

Gamgarn Somprasertsri, Napassakorn Mahattanateeranan

 

© 2025 Napassakorn Mahattanateeranan, published by UIKTEN. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. (CC BY-NC-ND 4.0)

 

Citation Information: TEM Journal. Volume 14, Issue 4, Pages 3155-3164, ISSN 2217-8309, DOI: 10.18421/TEM144-24, November 2025.

 

Received: 14 July 2025.
Revised: 25 August 2025.
Accepted: 08 September 2025.
Published: 27 November 2025.

 

Abstract:

 

This research investigates the use of the Bidirectional Encoder Representations from Transformers (BERT) model combined with Word2Vec embeddings and Word Mover's Distance (WMD) to retrieve aspect-level opinions from hotel reviews. The research focuses on classifying subjective sentences and ranking relevant opinionated sentences based on specific aspects, such as room quality, location, staff, and food. The semantic similarity of Word2Vec embeddings and the BERT model's ability to understand context in both directions were used together to make opinion retrieval tasks more accurate. Experimental results using TripAdvisor hotel reviews confirm that the proposed approach outperforms the traditional vector space model in terms of precision, recall, and F1-score. This study shows how advanced natural language processing techniques can be used to make opinion retrieval systems better. It also suggests areas that could be studied further in the future, like combining different types of data and finding the best retrieval methods for different types of review domains.

 

Keywords – Opinion retrieval, Hotel reviews, BERT model, Word embedding, Word Mover's Distance.

 

-----------------------------------------------------------------------------------------------------------

Full text PDF >  

-----------------------------------------------------------------------------------------------------------

 


Copyright © 2025 UIKTEN
Copyright licence: All articles are licenced via Creative Commons CC BY-NC-ND 4.0 licence