텍스트 마이닝을 활용한 음주 운전에 대한 미래 신호 탐색 - 대법원 양형기준 개정을 중심으로

This study aims to propose a futuristic methodology as a new approach to the problem of drunk driving and explore more effective ways to reduce drunk driving fatalities. Although there has been a significant decrease in drunk driving fatalities due to the strengthened punishment after the revision of the Yoon Changho Law, there is an opinion that simply strengthening the punishment is not enough to completely solve the problem of drunk driving due to the high recidivism rate and social controversy. Therefore, this study aims to analyze NAVER News data using text mining to detect risk factors and social signals related to drunk driving at an early stage and propose effective accident prevention policies. Data analysis using text mining is a useful method for early prediction and response to social problems by extracting meaningful information from news texts, which are unstructured data. In this study, we apply this methodology to the problem of drunk driving, with a focus on early detection of potential risk factors and, in particular, identifying signs of escalating blame or risk associated with drunk driving to guide policy improvements. Specifically, we explore the issue of 'dangerous driving manslaughter' through media coverage of controversial drunk driving cases, and use this to explore the need for strengthening drunk driving punishment policies as well as ways to raise public awareness to prevent drunk driving.