EUVIMEDEUVIMED entdecken
Crossref

Study on performance optimization of a voice phishing consultation chatbot using retrieval-augmented generation

Dohyeon Kong, Jiho Jung, Juhyun Choi, Kyungjong Kim

Journal of Data Forensics Research · 31.12.2025 · 2 (2) · 161-174

DOI: 10.12972/jdfr.2025.2.2.3 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Data Forensics Research, online erschienen am 31.12.2025. Verfasst von Dohyeon Kong und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

As voice-phishing methods become more sophisticated, the need for artificial intelligence counseling chatbots capable of effective responses has increased. However, existing large language model (LLM)-based chatbots have limitations in counseling applications due to the “hallucination” problem, where false information is presented as fact. To develop a reliable counseling system, this study applied retrieval-augmented generation (RAG) technology and proposed an optimization methodology to improve chatbot performance. First, data augmentation through sentence paraphrasing was performed to improve response quality. Second, five state-of-the-art embedding models were compared to identify the optimal model for search accuracy, resulting in the adoption of “multilingual-e5-large.” Finally, prompt structure and reference document ordering, which are key factors in RAG performance, were optimized through experimentation. Evaluation results show that the proposed optimized RAG model significantly outperforms the stand-alone LLM. This study demonstrates practical possibilities for building high-performance AI counseling systems applicable to real-world voice-phishing responsescenarios.

Zitieren

APA

Kong, D., Jung, J., Choi, J., & Kim, K. (2025). Study on performance optimization of a voice phishing consultation chatbot using retrieval-augmented generation. Journal of Data Forensics Research, 2(2), 161-174. https://doi.org/10.12972/jdfr.2025.2.2.3

Vancouver

Kong D, Jung J, Choi J, Kim K. Study on performance optimization of a voice phishing consultation chatbot using retrieval-augmented generation. Journal of Data Forensics Research. 2025;2(2):161-174. doi:10.12972/jdfr.2025.2.2.3

BibTeX
@article{Kong2025,
  author = {Kong, Dohyeon and Jung, Jiho and Choi, Juhyun and Kim, Kyungjong},
  title = {Study on performance optimization of a voice phishing consultation chatbot using retrieval-augmented generation},
  journal = {Journal of Data Forensics Research},
  year = {2025},
  volume = {2},
  number = {2},
  pages = {161--174},
  doi = {10.12972/jdfr.2025.2.2.3}
}
Kostenlos starten →