Development of a Legal Precedent Analysis System based on LLM
JOIV : International Journal on Informatics Visualization · 31.07.2026 · 10 (4) · 1824
DOI: 10.62527/joiv.10.4.5959 ↗Kurz gefasst
Fachartikel in der Zeitschrift JOIV : International Journal on Informatics Visualization, online erschienen am 31.07.2026. Verfasst von Hyo-joo Woo und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
School violence incidents constitute complex issues that require both legal and administrative judgment. Nevertheless, it is difficult for non-experts such as students, parents, and teachers to fully understand the criteria and procedures for disciplinary measures. Aware of this problem, this study designed and implemented a web service to support the handling of school violence cases based on legal precedent analysis. The proposed system structures users’ natural language descriptions of incidents through a conversational information collection module and analyzes them by quantifying key decision factors such as severity, persistence, intentionality, remorse, and the possibility of reconciliation. In addition, the system provides analytical results regarding the expected level of disciplinary measures through a dual prediction architecture that combines an LLM-based contextual analysis model with a deep learning classification model. By integrating these two approaches, the system improves the reliability and stability of prediction results while reflecting both contextual understanding and statistical patterns derived from past decisions. Furthermore, a BM25-based similar case matching function utilizing administrative adjudication precedents was implemented to provide evidence-based information linking prediction results to actual cases. This function allows users to review comparable precedents and understand how similar situations were handled in practice. The performance evaluation was conducted by comparing the system’s predictions with the judgments of 16 members of the School Violence Countermeasure Committee and by verifying consistency through stylistic variation cases. The results indicate that the proposed system can support users in understanding disciplinary criteria and procedures more effectively.
Zitieren
APA
Woo, H., Park, J., Lee, S., & Jo, J. (2026). Development of a Legal Precedent Analysis System based on LLM. JOIV : International Journal on Informatics Visualization, 10(4), 1824. https://doi.org/10.62527/joiv.10.4.5959
Vancouver
Woo H, Park J, Lee S, Jo J. Development of a Legal Precedent Analysis System based on LLM. JOIV : International Journal on Informatics Visualization. 2026;10(4):1824. doi:10.62527/joiv.10.4.5959
BibTeX
@article{Woo2026,
author = {Woo, Hyo-joo and Park, Jung-bin and Lee, Soo and Jo, Jaechoon},
title = {Development of a Legal Precedent Analysis System based on LLM},
journal = {JOIV : International Journal on Informatics Visualization},
year = {2026},
volume = {10},
number = {4},
pages = {1824},
doi = {10.62527/joiv.10.4.5959}
}
EUVIMED