PURPOSES : The primary objective of this study is to analyze the relationship between the factors that affect traffic incident duration in the mainline, tunnel, and ramp segments of an expressway. In addition, this study derived the most suitable statistical prediction model based on various incident duration distributions. METHODS : South Korean expressway crash data for 11 years, from 2011 to 2021, were analyzed. The incident durations on the mainline, tunnel, and ramp segments were selected using the accelerated failure time model, which is a parametric survival analysis approach. RESULTS : The mainline segment showed that the incident duration increased during accidents, including guard pipe collisions, multivehicle collisions, and snowfall. In particular, collisions in a tunnel with shoulder facilities increase the incident duration, while decreasing the time in the ramp segment. CONCLUSIONS : The incident duration model for each segment type yielded the most accurate results when applying a log-logistic distribution.

1. 서론
2. 기존문헌 고찰
3. 생존 모형(Survival model)
4. 데이터 수집 및 전처리
5. 분석결과
    5.1. AFT 모형의 분포 선정
    5.2. 사고처리시간 영향요인
    5.3. 모형의 정확도 검정
6. 결론
감사의 글

• 김강현(계명대학교 연구원) | Kim KangHyun (Department of Transportation, Keimyung University, 1095 Dalgubeol-daero, Dalseo-gu, Daegu 42601, Korea)
• 홍정열(정회원 · 계명대학교 조교수) | Hong JungYeol (Department of Transportation, Keimyung University, 1095 Dalgubeol-daero, Dalseo-gu, Daegu 42601, Korea) Corresponding Author