Explainable Multilingual Aspect-Based Sentiment Analysis for Tourism Using SHAP and LIME

Basworo Ardi Pramono; Rahmat Gernowo; Aghus Sofwan

doi:10.48084/etasr.18774

Authors

Basworo Ardi Pramono Doctor of Information Systems, Diponegoro University, Semarang, Indonesia | Informatics Engineering Department, Faculty of Information and Communication Technology, Semarang University, Semarang, Indonesia
Rahmat Gernowo Doctor of Information Systems, Diponegoro University, Semarang, Indonesia https://orcid.org/0000-0002-2409-7295
Aghus Sofwan Department of Electrical Engineering, Diponegoro University, Semarang, Indonesia https://orcid.org/0009-0004-8666-2968

Volume: 16 | Issue: 3 | Pages: 37077-37084 | June 2026 | https://doi.org/10.48084/etasr.18774

Received: 16 March 2026 | Revised: 23 April 2026 | Accepted: 11 May 2026 | Online: 6 June 2026

Corresponding author: Basworo Ardi Pramono

Abstract

Online tourism analytics increasingly relies on Aspect-Based Sentiment Analysis (ABSA) to extract fine-grained visitor perceptions; however, prior tourism ABSA studies often emphasize predictive performance while providing limited and rarely quantified evidence on explanation reliability. To address this gap, this study proposes an explainable multilingual ABSA framework for tourism reviews that combines one-vs-rest Logistic Regression (LR) with linear coefficients, SHAP, and LIME, and augments them with a quantitative trustworthiness evaluation. Experiments were conducted on a bilingual corpus of 2,891 Indonesian and English Google Reviews collected from 10 tourist destinations in Central Java and annotated into 9 multi-label classes derived from the dimensions of attractions, amenities, and accessibility, and their sentiment polarities. The selected model achieved a Macro-F1 of 0.4586, a Hamming loss of 0.1505, and an exact match of 0.2512. The global explanation analysis showed substantial agreement between the LR coefficients and SHAP rankings, with overlap@10 generally ranging from 0.70 to 0.80 across most labels. Eraser-based evaluation at = 10 preserved predictions for 0.72–1.00 of cases, indicating strong fidelity of influential features. At the local level, SHAP and LIME consistently highlighted salient tokens associated with both correct and incorrect predictions, while sanity checks showed sharp degradation under model randomization, confirming that the explanations were tied to learned model parameters rather than superficial artifacts. These findings demonstrate that multilingual tourism ABSA can be made both interpretable and quantitatively auditable, thereby providing a transparent analytical basis for tourism service evaluation, destination management, and future decision-support applications.

Keywords:

aspect-based sentiment analysis, explainable artificial intelligence, tourism analytics, SHAP, LIME

References

[1] M. Agua, N. Antonio, M. P. Carrasco, and C. Rassal, "Large Language Models Powered Aspect-Based Sentiment Analysis for Enhanced Customer Insights," Tourism & Management Studies, vol. 21, no. 1, pp. 1–19, Jan. 2025.

[2] A. Jain, A. Bansal, and S. Tomar, "Aspect-Based Sentiment Analysis of Online Reviews for Business Intelligence:," International Journal of Information Technologies and Systems Approach, vol. 15, no. 3, pp. 1–21, Aug. 2022.

[3] S. O. E. Putri, A. A. Arifiyanti, and A. R. E. Najaf, "Convolutional Neural Network Approach for Aspect-Based Sentiment Analysis of Tourism Reviews," bit-Tech, vol. 8, no. 1, pp. 448–458, Aug. 2025.

[4] A. Murzakhmetov, M. Satymbekov, A. Bapanov, and N. Beisov, "Sentiment Analysis of Tourist Reviews About Kazakhstan Using a Hybrid Stacking Ensemble Approach," Computation, vol. 13, no. 10, Oct. 2025, Art. no. 240.

[5] A. Alsehaimi, A. Babour, and D. Alahmadi, "Toward Transparent Modeling: A Scoping Review of Explainability for Arabic Sentiment Analysis," Applied Sciences, vol. 15, no. 19, Oct. 2025, Art. no. 10659.

[6] A. A. Maruf, F. Khanam, Md. M. Haque, Z. M. Jiyad, M. F. Mridha, and Z. Aung, "Challenges and Opportunities of Text-Based Emotion Detection: A Survey," IEEE Access, vol. 12, pp. 18416–18450, 2024.

[7] M. R. A. Yudianto, P. Sukmasetya, R. A. Hasani, and Maimunah, "Aspect-Based Sentiment Analysis of Borobudur Temple Reviews Use Support Vector Machine Algorithm," E3S Web of Conferences, vol. 500, 2024, Art. no. 01005.

[8] G. I. Bhaskara, I. G. A. Sastrawan, and I. G. B. A. Yudiastina, "Sentiment and Sunsets: Analysing Online Reviews of Kuta Beach in Bali," E-Journal of Tourism, Mar. 2024, Art. no. 76.

[9] S. Rajarajeswari, D. Ashwin, N. S. Kumar, R. Vishnal, and N. Vishwa, "Online Review Sentimental Analysis," International Journal of Innovative Science and Research Technology, pp. 2152–2155, May 2025.

[10] X. Chen, H. Xie, S. J. Qin, Y. Chai, X. Tao, and F. L. Wang, "Cognitive-Inspired Deep Learning Models for Aspect-Based Sentiment Analysis: A Retrospective Overview and Bibliometric Analysis," Cognitive Computation, vol. 16, no. 6, pp. 3518–3556, Nov. 2024.

[11] Y. Zhao, J. Zhang, Y. Tong, Z. Li, X. Yu, and S. Tsai, "Design of an Enterprise Public Opinion Monitoring System Based on Natural Language Processing: Sentiment Analysis and Management of Public Opinion Data," Journal of Global Information Management, vol. 33, no. 1, pp. 1–35, June 2025.

[12] A. Orive, A. Agirre, H. L. Truong, I. Sarachaga, and M. Marcos, "Quality of Service Aware Orchestration for Cloud–Edge Continuum Applications," Sensors, vol. 22, no. 5, Feb. 2022, Art. no. 1755.

[13] F. Tanveer et al., "Balancing privacy and performance in healthcare: A federated learning framework for sensitive data," Digital Health, vol. 11, May 2025, Art. no. 20552076251381769.

[14] F. Bodria, F. Giannotti, R. Guidotti, F. Naretto, D. Pedreschi, and S. Rinzivillo, "Benchmarking and survey of explanation methods for black box models," Data Mining and Knowledge Discovery, vol. 37, no. 5, pp. 1719–1778, Sept. 2023.

[15] Z. Zhang, H. A. Hamadi, E. Damiani, C. Y. Yeun, and F. Taher, "Explainable Artificial Intelligence Applications in Cyber Security: State-of-the-Art in Research," IEEE Access, vol. 10, pp. 93104–93139, 2022.

[16] S. Hameed, M. Nauman, N. Akhtar, M. A. B. Fayyaz, and R. Nawaz, "Explainable AI-driven depression detection from social media using natural language processing and black box machine learning models," Frontiers in Artificial Intelligence, vol. 8, Sept. 2025, Art. no. 1627078.

[17] S. Khanapur, J. S. Nayak, B. S. Rajeshwari, M. Namratha, C. B. Bharadwaj, and R. Bhardwaj, "SHAP-Based Explainability for Local and Global Insights in Alzheimer’s Detection," Engineering, Technology & Applied Science Research, vol. 16, no. 1, pp. 30940–30947, Feb. 2026.

[18] A. Adak, B. Pradhan, and N. Shukla, "Sentiment Analysis of Customer Reviews of Food Delivery Services Using Deep Learning and Explainable Artificial Intelligence: Systematic Review," Foods, vol. 11, no. 10, May 2022, Art. no. 1500.

[19] E. I. Setiawan, F. Ferry, J. Santoso, S. Sumpeno, K. Fujisawa, and M. H. Purnomo, "Bidirectional GRU for Targeted Aspect-Based Sentiment Analysis Based on Character-Enhanced Token-Embedding and Multi-Level Attention," International Journal of Intelligent Engineering and Systems, vol. 13, no. 5, pp. 392–407, Oct. 2020.

[20] S. Mirzaei, H. Mao, R. R. O. Al-Nima, and W. L. Woo, "Explainable AI Evaluation: A Top-Down Approach for Selecting Optimal Explanations for Black Box Models," Information, vol. 15, no. 1, Dec. 2023, Art. no. 4.

[21] E. Albini, A. Rago, P. Baroni, and F. Toni, "Achieving descriptive accuracy in explanations via argumentation: The case of probabilistic classifiers," Frontiers in Artificial Intelligence, vol. 6, Apr. 2023, Art. no. 1099407.

[22] B. A. Pramono, "Dual-Annotated Central Java Tourism Review Dataset: 10 Tourist Destinations with Annotator A and Annotator B." Zenodo, Mar. 10, 2026.

[23] M. Danilevsky, S. Dhanorkar, Y. Li, L. Popa, K. Qian, and A. Xu, "Explainability for Natural Language Processing," in Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, Aug. 2021, pp. 4033–4034.

[24] F. Stoehr et al., "Natural language processing for automatic evaluation of free-text answers — a feasibility study based on the European Diploma in Radiology examination," Insights into Imaging, vol. 14, no. 1, Sept. 2023, Art. no. 150.

[25] E. Balkir, S. Kiritchenko, I. Nejadgholi, and K. Fraser, "Challenges in Applying Explainability Methods to Improve the Fairness of NLP Models," in Proceedings of the 2nd Workshop on Trustworthy Natural Language Processing (TrustNLP 2022), 2022, pp. 80–92.

[26] J. Yuan, J. Vig, and N. Rajani, "iSEA: An Interactive Pipeline for Semantic Error Analysis of NLP Models," in 27th International Conference on Intelligent User Interfaces, Mar. 2022, pp. 878–888.

[27] S. Gurrapu, A. Kulkarni, L. Huang, I. Lourentzou, and F. A. Batarseh, "Rationalization for explainable NLP: a survey," Frontiers in Artificial Intelligence, vol. 6, Sept. 2023, Art. no. 1225093.