Aspect-Level Sentiment Analysis Using WangchanBERTa for Fine-Grained Service Insight Extraction in Hotel Reviews
Main Article Content
Abstract
Online customer reviews on hotel booking platforms significantly influence consumer decisions and the reputation of SME hotels in Thailand. The critical challenge for hoteliers lies not in data scarcity but in efficiently extracting actionable insights from unstructured Thai-language reviews. Thus, this research presents an automated sentiment analysis and strategic insight generation system for hotels, leveraging WangchanBERTa, a Thai-specific deep learning model. The system comprises two phases. Phase 1 develops a sentiment classification model using 10,040 Thai hotel reviews collected from Agoda, Booking.com, Traveloka, and Trip.com, categorizing sentiments into positive and negative classes while identifying key topics such as pricing, service quality, and cleanliness. Phase 2 extracts granular aspect-level insights across 11 service dimensions to detect nuanced patterns, such as customers being satisfied with service but dissatisfied with pricing. Experimental results demonstrate robust model performance with an accuracy of 92.62% and a macro F1-score of 94.86% for overall sentiment classification. For aspect-based sentiment analysis, the system achieved 98.66% accuracy with macro precision of 93.62%, recall of 94.62%, and F1-score of 94.22%, alongside effective real-world insight extraction capabilities. This framework enables hoteliers to deeply understand customer voices, transform data into actionable business strategies, and enhance competitive positioning in Thailand's tourism sector.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
This journal provides immediate open access to its content on the principle that making research freely available to the public supports a greater global exchange of knowledge.
- Creative Commons Copyright License
The journal allows readers to download and share all published articles as long as they properly cite such articles; however, they cannot change them or use them commercially. This is classified as CC BY-NC-ND for the creative commons license.
- Retention of Copyright and Publishing Rights
The journal allows the authors of the published articles to hold copyrights and publishing rights without restrictions.
References
L. Lowphansirikul, C. Polpanumas, N. Jantrakulchai, and S. Nutanong, "WangchanBERTa: Pretraining transformer-based Thai language models," arXiv preprint arXiv:2101.09635, 2021, doi: 10.48550/arXiv.2101.09635.
T. Wolf et al., "Transformers: State-of-the-art natural language processing," in Proc. Empirical Methods in Natural Language Processing: System Demonstrations (EMNLP 2020), Online, Oct. 2020, pp. 38–45, doi: 10.18653/v1/2020.emnlp-demos.6.
C. Chantrapornchai and A. Tunsakul, "Information extraction tasks based on BERT and SpaCy on tourism domain," ECTI Trans. Comput. Inf. Technol., vol. 15, no. 1, pp. 108–122, 2021, doi: 10.37936/ecti-cit.2021151.228621.
N. Khamphakdee and P. Seresangtakul, "Sentiment analysis for Thai language in hotel domain using machine learning algorithms," Acta Informatica Pragensia, vol. 10, no. 2, pp. 155–171, 2021, doi: 10.18267/j.aip.155.
N. Khamphakdee and P. Seresangtakul, “An Efficient Deep Learning for Thai Sentiment Analysis,” Data, vol. 8, no. 5, art. 90, 2023, doi: 10.3390/data8050090.
K. Namee, J. Polpinij, and B. Luaphol, "A hybrid approach for aspect-based sentiment analysis: A case study of hotel reviews," Curr. Appl. Sci. Technol., vol. 23, no. 2, 2023, doi: 10.55003/cast.2022.02.23.008.
Y. Bai et al., "Is compound aspect-based sentiment analysis addressed by LLMs?" in Findings Assoc. Comput. Linguistics: EMNLP 2024, Miami, FL, USA, Nov. 2024, pp. 7836–7861, doi: 10.18653/v1/2024.findings emnlp.460.
T. Lapchaicharoenkit and P. Vateekeul, "Machine reading comprehension using multi-passage BERT with Dice loss on Thai corpus," ECTI Trans. Comput. Inf. Technol., vol. 16, no. 2, pp. 125–134, 2022, doi: 10.37936/ecti-cit.2022162.247799.
Y.-C. Hua, Z. Zhou, Q. Han, Y. Li, R. He, and J. Zhou, "A systematic review of aspect-based sentiment analysis," Artif. Intell. Rev., vol. 57, art. 296, 2024, doi: 10.1007/s10462-024-10906-z.
I. Nawawi, K. F. Ilmawan, M. R. Maarif, and M. Syafrudin, "Exploring tourist experience through online reviews using aspect-based sentiment analysis with zero shot learning for hospitality service enhancement," Information, vol. 15, no. 8, art. 499, 2024, doi: 10.3390/info15080499.
G. Kontonatsios, J. Clive, G. Harrison, T. Metcalfe, P. Sliwiak, H. Tahir, and A. Ghose, "An aspect-based sentiment analysis dataset of user reviews," Neurocomputing, vol. 562, art. 126867, 2023, doi: 10.1016/j.neucom.2023.126867.
A. Ameur, S. Hamdi, and S. Ben Yahia, “Sentiment analysis for hotel reviews: A systematic literature review,” in Companion Proc. 35th Int. Conf. Advanced Information Systems Engineering (CAiSE), Zaragoza, Spain, Jun. 2023, pp. 127–142, doi: 10.1145/3605152.
K. Suraratchai and S. Phoomvuthisarn, "Thai language sentiment analysis with a hybrid method on WangchanBERTa-CNN-BiLSTM," J. Inf. Sci. Technol., vol. 14, no. 2, pp. 1–11, 2024.
W. Bunyatisai, S. Plubin, K. Jiamwattanapong and B. Plubin, “Improving Thai sentiment analysis accuracy with emoji classification by deep learning and stacking models: A case study of hotel reviews,” Pakistan J. Life Soc. Sci., vol. 23, no. 1, pp. 65–77, 2025, doi: 10.57239/PJLSS 2025-23.1.007.
N. Jensirisak, D. Tasachan, and P. Wanchai, "Aspect based sentiment analysis for Thai language from consumer reviews towards smartphone," J. Sci. Ladkrabang, vol. 34, no. 1, pp. 20–42, 2025.
R. Wanasukapunt and S. Phimoltares, “Classification of abusive Thai language content in social media using deep learning,” in Proc. 2021 18th Int. Joint Conf. Computer Science and Software Engineering (JCSSE), 2021, pp. 1–6. doi: 10.1109/JCSSE53117.2021.9503294.
M. Nokkaew, et al., "Analyzing online public opinion on Thailand–China high-speed train and Laos–China railway mega-projects using advanced machine learning for sentiment analysis," Soc. Netw. Anal. Min., vol. 14, no. 1, art. 15, 2024, doi: 10.1007/s13278-023-01168-8.
Y. Wen, Y. Liang and X. Zhu, “Sentiment analysis of hotel online reviews using the BERT model and ERNIE model — Data from China,” PLoS ONE, vol. 18, no. 3, art. e0275382, Mar. 2023, doi:10.1371/journal.pone.0275382.
M. Thomson, H. Murfi and G. Ardaneswari, “BERT Based Hybrid Deep Learning with Text Augmentation for Sentiment Analysis of Indonesian Hotel Reviews,” in Proc. 12th Int. Conf. Data Science, Technology and Applications (DATA 2023), Rome, Italy, Jul. 2023, pp. 468–473, doi: 10.5220/0012127400003541.
S. C. Sahin and C. Eyupoglu, "Aspect-based sentiment analysis for hospitality industry applications: A systematic literature review," Appl. Comput. Syst., vol. 30, no. 1, pp. 53–67, 2025, doi: 10.2478/acss-2025-0007.
K. I. Roumeliotis, N. D. Tselikas and D. K. Nasiopoulos, “Leveraging Large Language Models in Tourism: A Comparative Study of the Latest GPT Omni Models and BERT NLP for Customer Review Classification and Sentiment Analysis,” Information, vol. 15, no. 12, art. 792, 2024, doi: 10.3390/info15120792.
M. Nokkaew et al., "Hidden emotional trends on social media regarding the Thailand–China high-speed railway project: A deep learning approach with ChatGPT integration," Soc. Netw. Anal. Min., vol. 14, no. 1, art. 175, 2024, doi: 10.1007/s13278-024-01340-8.
A. Jazuli, Widowati and R. Kusumaningrum, “Optimizing Aspect-Based Sentiment Analysis Using BERT,” Appl. Sci., vol. 15, no. 1, art. 172, 2025, doi: 10.3390/app15010172.
I. A. Kandhro, F. Ali, M. Uddin, A. Kehar, and S. Manickam, "Exploring aspect-based sentiment analysis: An in-depth review of current methods and prospects for advancement," Knowl. Inf. Syst., vol. 66, pp. 3639–3669, 2024, doi: 10.1007/s10115-024-02104-8.
J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, "BERT: Pre-training of deep bidirectional transformers for language understanding," in Proc. Conf. North American Chapter Assoc. Comput. Linguistics: Human Language Technologies (NAACL-HLT), Minneapolis, MN, USA, Jun. 2019, pp. 4171–4186, doi: 10.18653/v1/N19-1423.
J. Geng, F. Cai, Y. Wang, H. Koeppl, P. Nakov, and I. Gurevych, "A survey of confidence estimation and calibration in large language models," in Proc. 2024 Conf. North American Chapter Assoc. Comput. Linguistics: Human Language Technologies (NAACL-HLT), Mexico City, Mexico, Jun. 2024, pp. 6577–6595, doi: 10.18653/v1/2024.naacl-long.366.
Y. Cao, H. Mozannar, L. Feng, H. Wei, and B. An, "In defense of softmax parametrization for calibrated and consistent learning to defer," in Proc. 37th Annu. Conf. Neural Information Processing Systems (NeurIPS), New Orleans, LA, USA, Dec. 2023, pp. 38485–38503.
F. Adjed, S. Herbin, and A. LeCoz, "Confidence calibration of classifiers with many classes," in Proc. 38th Annu. Conf. Neural Information Processing Systems (NeurIPS), Vancouver, Canada, Dec. 2024, doi: 10.52202/079017-2469.
W. Phatthiyaphaibun et al., "PyThaiNLP: Thai natural language processing in Python," in Proc. 3rd Workshop Natural Language Processing Open Source Software (NLP-OSS 2023), Singapore, Dec. 2023, pp. 21–30, doi: 10.18653/v1/2023.nlposs-1.4.
P. Phatthiyaphaibun, C. Polpanumas, and T. Theeramunkong, "Deep transformer-based models for Thai sentiment and emotion classification," ICT Express, vol. 9, no. 2, pp. 221–230, 2023, doi: 10.1016/j.icte.2022.12.003.
