PURPOSES : This study aims to conduct a sensitivity analysis to determine the major factors affecting traffic accidents involving elderly pedestrians.
METHODS : In this study, a regression tree model was built based on a non-parametric statistical model using data on traffic accidents involving elderly pedestrians. Using this model, we analyzed the degree of change in the probability of pedestrian fatalities.
RESULTS : Results of the model analysis show that the first major factor combination affecting traffic accidents involving elderly pedestrians is speeding, night time, and road markers. The second combination is night time and arterial roads (national and local highways). The last combination that may lead to such accidents is heavy vehicles and federally funded local highways.
CONCLUSIONS : Preventive measures, such as speed control, proper lighting, median strips, designation of pedestrian protection zones, and guidance of detours, are necessary to manage high-risk combinations causing accidents of the elderly.

ABSTRACT
1. 서론
    1.1. 연구의 필요성 및 목적
    1.2. 연구의 내용
2. 선행연구 고찰
    2.1. 고령 보행자 교통사고 요인 분석 선행연구 고찰
    2.2. 선행연구의 한계와 본 연구의 차별성
    2.3. Regression Tree 모형 이론 고찰
3. 분석방법론 정립 및 자료수집
    3.1. 분석방법론 정립
    3.2. 분석 자료수집
    3.3. 수집 자료의 기초통계분석
4. 모형 구축
    4.1. 종속변수 설정 및 독립변수 선택 방법
    4.2. 독립변수 선택 방법별 선정 결과 비교
    4.3. 독립변수 선정
    4.4. 모형 결과 해석 방법
5. 주요 사고요인 조합과 민감도 분석
    5.1. 주요 사고요인 조합1(과속)
    5.2. 주요 사고요인 조합2(주야 등)
    5.3. 지방지역도로 고령 보행자 교통사고 조합 유형화
    5.4. 사고요인 조합에 따른 민감도 분석
6. 결론
REFERENCES

• 이동윤(한국교통연구원 도로교통연구본부 연구원) | Lee Dong-youn
• 박호철(명지대학교 교통공학과 교수) | Park Ho-chul Corresponding author
• 우승국(한국교통연구원 도로교통연구본부 교통안전·방재연구센터장) | Wu Seung-kook
• 심재익(한국교통연구원 도로교통연구본부 교통안전·방재연구센터 선임연구위원) | Sim Jae-ik
• 정가형(수원시정연구원 도시공간연구실 연구원) | Jeong Ga-hyung
• 이의은(명지대학교 교통공학과 교수) | Lee Eui-eun