Natural Language Processing-based Judgement Prediction System for Road Traffic Accidents: Focused on Text Information for Traffic Accident Situations

Hyunsik Min; Junyeong Yun; Byeongjoon Noh

doi:10.7470/jkst.2024.42.4.385

All Issue

2024 Vol.42, Issue 4 Next Page

Article

Natural Language Processing-based Judgement Prediction System for Road Traffic Accidents: Focused on Text Information for Traffic Accident Situations 자연어처리 기반의 도로교통사고 판결예측 시스템: 교통사고상황 텍스트 정보 활용을 중심으로

31 August 2024. pp. 385-397

PDF XML

Abstract

Traffic accident scenarios are inherently complex, requiring substantial time and effort to analyze textual information pertaining to the causes, circumstances, and outcomes for legal adjudication. Such complexity also introduces a potential for human errors. Therefore, it is necessary to devise an efficient system for analyzing such textual data. In this study, we propose a natural language processing (NLP)-based system that can predict judicial verdicts in traffic accident cases. The proposed system leverages NLP techniques to extract and analyze information related to traffic laws and regulations from textual descriptions of traffic accident scenarios, subsequently predicting sentencing information such as probation status and fines. The proposed system mainly consists of three parts: 1) data collection and preprocessing, 2) construction of a word dictionary and embedding, and 3) verdict prediction. First, traffic accident-related judgments are collected from various legal databases. Subsequently, key information such as imprisonment terms, probation status, and fines is extracted from the collected judgments, and a domain-specific word dictionary tailored to traffic accident cases is built using embedding techniques. Finally, the legal verdicts are predicted using a deep learning model. In our experiment, we validated the feasibility and applicability of the proposed system by implementing and applying it to approximately 37,406 traffic accident-related judgments, resulting in accuracies of 96.13%, 94.85%, and 95.49% for imprisonment, probation status, and fines, respectively.

Keywords

case law prediction

natural language processing

road traffic accidents

traffic accident judgement prediction system

traffic accident situation information

교통사고 상황은 대체로 매우 복잡하여 법적 판단을 위해 사고의 원인, 경위, 결과 등의 텍스트 정보를 분석하는 것은 높은 시간 비용이 요구되고, 오류의 가능성을 배제할 수 없기에 이러한 텍스트 정보를 효율적으로 분석하기 위한 시스템이 필요하다. 이러한 문제를 해결하기 위해서 본 연구에서는 교통사고 상황에 대한 법원 판결을 예측하기 위한 자연어 처리 시스템을 제안한다. 제안하는 시스템은 자연어 처리 방법론을 활용하여 교통사고 상황에 관한 텍스트 정보로부터 도로교통법 등 교통법규 관련 정보를 분석하여 해당 상황에 대한 형량 정보(집행유예 여부 및 벌금액 등)를 예측한다. 제안하는 시스템은 크게 1) 데이터 수집 및 전처리, 2) 단어 사전 구축 및 임베딩, 3) 판결 예측의 세 과정으로 구분된다. 먼저, 데이터 수집 및 전처리 과정에서는 다양한 법률 데이터베이스로부터 교통사고 관련 판결문을 수집 및 전처리한다. 다음으로 징역 및 집행유예 여부, 벌금액 등과 같은 판결문 내의 중요 정보를 추출하고, 임베딩을 통해 교통사고 판례에 특화된 단어 사전을 구축한다. 마지막으로 판결 예측 모델에서는 딥러닝 모델을 통해 판결문으로부터 형량을 예측한다. 본 연구의 실험에서는 약 37,406건의 도로교통법 관련 판결 데이터를 활용하여 징역, 집행유예, 벌금 결과를 예측하여 제안하는 시스템의 실현가능성을 검증하였으며 그 결과는 각각 96.13%, 94.85%, 95.49%의 정확도로 나타났다.

키워드

판결예측

자연어처리

도로교통사고

교통사고 판결예측 시스템

교통사고상황 정보

References

Bojanowski P., Grave E., Joulin A., Mikolov T. (2017), Enriching Word Vectors with Subword Information, Transactions of the Association for Computational Linguistics, 5, 135-146.

10.1162/tacl_a_00051

Byun J. U., Kim H. S., Park M. R., Shin J. W. (2018), Court Decision Estimation using Machine Learning Algorithms, Journal of Korean Criminological Association, 12(1), 3-17.

10.29095/JKCA.12.1.1

Choung W. (2023), A Study on the Legislative Issues of Artificial Intelligence, The Institute of Legal Studies, Kyunghee University, South Korea, 58(4), 117-144.

10.15539/KHLJ.58.4.5

European Convention on Human Rights, https://www.echr.coe.int/european-convention-on-human-rights, 2024.1.29.

Goldberg D. M. (2022), Characterizing Accident Narratives with Word Embeddings: Improving Accuracy, Richness, and Generalizability, Journal of Safety Research, 80, 441-455.

10.1016/j.jsr.2021.12.02435249625

Graham S. G., Soltani H., Isiaq O. (2023), Natural Language Processing for Legal Document Review: Categorising Deontic Modalities in Contracts, Artificial Intelligence and Law, 1-22.

10.1007/s10506-023-09379-2

Hwang S. H. (2009), The Idea and Reality of the Civil Legal Aid System in Korean Society, Legal Research Institute KMU Seoul Korea, 21(2), 247-280.

10.17251/legal.2009.21.2.247

Kim B. P. (2021), Predictive Justice Technology and Its Limitations, Korean Society of Law, 182-2, 89-117.

10.29305/tj.2021.02.182.2.89

Kim C. B. (2022a), Deep Learning Model for Prediction Rate Improvement of Weather Data using Parallel Merge Structure, Journal of Korean Institute of Information Technology, 20(4), 131-140.

10.14801/jkiit.2022.20.4.131

Kim Y. H. (2022b), Analysis and Legal Strategy on the Traffic Law in the Advent of the 4th Industrial Revolution and Road Environment Change, The Journal of Legal Studies, 30(2), 163-190.

10.35223/GNULAW.30.2.7

Kim C. H., Lee S. J., Kim J. H. (2022), Exploring the Features of Aggressive Driving and Road Rage by Analyzing Judicial Precedents Based on Topic Modeling, J. Korean Soc. Transp., 40(5), Korean Society of Transportation, 683-699.

10.7470/jkst.2022.40.5.683

LBox, https://lbox.kr, 2023.12.21.

Lecun Y., Bottou L., Bengio Y., Haffner P. (1998), Gradient-based Learning Applied to Document Recognition, Proceedings of the IEEE, 86(11), 2278-2324.

10.1109/5.726791

Lee D. B., Choi S. P. (2019), Comparative Analysis of Various Korean Morpheme Embedding Models using Massive Textual Resources, Korean Institute of Information Scientists and Engineers, 46(5), 413-418.

10.5626/JOK.2019.46.5.413

Lee Y. G. (2023), Performance Comparison of Automatic Classification Using Word Embeddings of Book Titles, Korean Society for Information Management, 40(4), 307-327.

Medvedeva M., Vols M., Wieling M. (2020), Using Machine Learning to Predict Decisions of the European Court of Human Rights, Artificial Intelligence and Law, 28(2), 237-266.

10.1007/s10506-019-09255-y

National Law Information Center (2024a, January 29), Road Traffic Act. Retrieved from https://law.go.kr/LSW/ eng/engLsSc.do?menuId=2&section=lawNm&query=Road+Traffic+Act&x=0&y=0#liBgcolor1

National Law Information Center (2024b, January 29), Traffic Accident Handling Act. Retrieved from https:// law.go.kr/LSW/eng/engLsSc.do?menuId=2&section=lawNm&query=Act+on+Special+Cases+concerning+the+Settlement+of+Traffic+Accidents&x=0&y=0#liBgcolor0

Noh B. J., Yeo H. S. (2021), SafetyCube: Framework for Potential Pedestrian Risk Analysis Using Multi-dimensional OLAP, Accident Analysis & Prevention, 155, 106104, ISSN 0001-4575.

10.1016/j.aap.2021.10610433819792

Pennington J., Socher R., Manning C. (2014), GloVe: Global Vectors for Word Representation, EMNLP 2014, 1532-1543.

10.3115/v1/D14-1162

Sentencing Commission (2023, December 21), Sentencing Guidelines for Traffic Crimes. Retrieved from https:// sc.scourt.go.kr/sc/krsc/criterion/criterion_35/traffic_change_01.jsp

Umer M., Imtiaz Z., Ahmad M., Nappi M., Medaglia C., Choi G. S. et al. (2023), Impact of Convolutional Neural Network and FastText Embedding on Text Classification, Multimedia Tools and Applications: An International Journal, 82(4), 5569-5585.

10.1007/s11042-022-13459-x

Yoon H. S. (2016), Study on the Problems and Solutions of the Act on Special Cases Concerning the Settlement of Traffic Accidents, Korean Police Studies Association, 15(4), 97-130.

Zahir J. (2023), Prediction of Court Decision from Arabic Documents Using Deep Learning, Expert Systems, 40(6), 1-16.

10.1111/exsy.13236

Information

Publisher :Korean Society of Transportation
Publisher(Ko) :대한교통학회
Journal Title :Journal of Korean Society of Transportation
Journal Title(Ko) :대한교통학회지
Volume : 42
No :4
Pages :385-397
Received Date : 2024-03-14
Revised Date : 2024-04-24
Accepted Date : 2024-08-05
DOI :https://doi.org/10.7470/jkst.2024.42.4.385

Journal of Korean Society of Transportation ISSN:1229-1366(Print) 2234-4217(Online) 대한교통학회지

All Issue