Localized Cebuano-English DRRM Chatbot via Retrieval-Augmented Generation and Large Language Models
DOI:
https://doi.org/10.5281/zenodo.21158345Keywords:
DRRM chatbot; Cebuano–English; Retrieval-Augmented Generation; Large Language Model; Technology Acceptance ModelAbstract
This study designed, developed, and evaluated a localized Cebuano-English Disaster Risk Reduction and Management (DRRM) chatbot utilizing Retrieval-Augmented Generation (RAG) and Large Language Models (LLMs) to meet residents' demands for clearer disaster communication regarding typhoons, floods, evacuation protocols, emergency contacts, go-bag preparation, and weather advisories.
The research employed a developmental-descriptive design. The developmental phase involved building a web-based chatbot prototype powered by an embedding-based RAG architecture. In this pipeline, verified DRRM knowledge entries were chunked, converted into vector embeddings, and stored in a repository for accurate semantic retrieval. The LLM generated localized Cebuano-English responses strictly grounded in the retrieved, validated content. To ensure information safety and prevent hallucination, a administrative ingestion module was integrated, requiring manual review and approval by MDRRMO officials prior to data indexation.
The descriptive phase evaluated baseline DRRM communication channels and system acceptability using the Technology Acceptance Model (TAM). Findings identified typhoons and floods as the primary hazards faced by residents, with radio, barangay announcements, MDRRMO/LGU alerts, and social media serving as main information sources. Respondents expressed a clear preference for Cebuano or bilingual Cebuano-English explanations—especially when clarifying technical disaster terminology. TAM evaluation results indicated exceptionally high user acceptability, achieving an overall grand mean of 4.45 ($SD$-equivalent high rating), demonstrating that users found the chatbot highly useful, intuitive, and effective as a supplementary DRRM communication tool.
The study concluded that the AI-driven, localized chatbot serves as a valuable supplementary tool for municipal and barangay-level DRRM communication, though it is designed to augment rather than replace official emergency mandates, advisories, or evacuation orders issued by authorized disaster management agencies.
References
Brown, J., & Smith, L. (2021). AI-driven career recommendations: A systematic review. International Journal of Career Development, 14(3), 125–140.
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D. M., Wu, J., Winter, C., & Amodei, D. (2020). Language models are few-shot learners (arXiv:2005.14165). arXiv. https://doi.org/10.48550/arXiv.2005.14165
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340. https://doi.org/10.2307/249008
Lewis, P. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. Advances in Neural Information Processing Systems, 33, 9459–9474.
National Disaster Risk Reduction and Management Council. (2020). National disaster risk reduction and management plan 2020–2030. Office of Civil Defense.
Philippine Atmospheric, Geophysical and Astronomical Services Administration. (n.d.). Flood information and advisories. Department of Science and Technology. https://www.pagasa.dost.gov.ph/
Republic Act No. 10121. (2010). Philippine Disaster Risk Reduction and Management Act of 2010. Official Gazette of the Republic of the Philippines. https://lawphil.net/statutes/repacts/ra2010/ra_10121_2010.html
United Nations Office for Disaster Risk Reduction. (2015). Sendai framework for disaster risk reduction 2015–2030. UNDRR. https://www.undrr.org/publication/sendai-framework-disaster-risk-reduction-2015-2030
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need (arXiv:1706.03762). arXiv. https://doi.org/10.48550/arXiv.1706.03762
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.