서브메뉴
검색
Crash Modification Factors, Predictive Hotspot Identification, and Treatment Selection in Texas and Peer States
Crash Modification Factors, Predictive Hotspot Identification, and Treatment Selection in Texas and Peer States
Detailed Information
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260311091547.5
- ISBN
- 9798270232641
- DDC
- 519.72
- 저자명
- Hasan, Manar
- 서명/저자
- Crash Modification Factors, Predictive Hotspot Identification, and Treatment Selection in Texas and Peer States / Manar Hasan
- 발행사항
- [Sl] : The University of Texas at Austin, 2025
- 형태사항
- 1 electronic resource (283 pages)
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-06, Section: A.
- 주기사항
- Advisors: Machemehl, Randy B. Committee members: Bhasin, Amit; Boyles, Stephen; Jiao, Junfeng.
- 학위논문주기
- - Ph.D. : The University of Texas at Austin, 2025.
- 초록/해제
- 요약Safety is crucial in transportation, with huge human and property costs associated with every crash. Proactive data-driven decision-making is key in determining safety treatments that will have maximum positive impact. Traditional processes rely on historical crash data, generalized Crash Modification Factors (CMFs), and reactive interventions, often limiting their effectiveness. This dissertation addresses three challenges in safety decision-making: (1) assessing the reliability of currently used CMFs in estimating crash reductions, (2) identifying future high-risk crash locations, and (3) analytically performing treatment selection to maximize safety benefits. First, this dissertation evaluates the accuracy of CMFs currently used by transportation agencies by comparing expected crash reductions with observed outcomes. The findings reveal that some CMFs closely align with real-world crash data, while others, may overestimate or underestimate safety benefits. Also, CMFs can vary significantly for different crash severity levels, so jurisdictions that use one generalized value are inaccurately estimating safety treatment effects. Second, the dissertation performs predictive hotspot identification, leveraging Poisson and Random Forest models to forecast crash-prone locations. The models demonstrate good predictive performance, with traffic volume (ADT), traffic factors (peak hour factor, directional factor), speed limits, and surface width emerging as key factors influencing crash risk. The results suggest that integrating statistical and machine learning analyses into safety planning can aid proactive crash prevention, shifting from traditional reactive approaches. Finally, the dissertation proposes a quantitative, data-driven framework for treatment selection. Using propensity score matching (PSM) and predictive modeling with many models (Difference-in-Differences, Linear Regression, Random Forest, Gradient Boosting, Poisson, Negative Binomial, XGBoost, SVR, and Neural Network), the research evaluates treatment effectiveness across different roadway conditions and treatments. The findings highlight models that performed better than others (Poisson, Negative Binomial, Gradient Boosting, Random Forest, and XGBoost), and also some conditions which make certain treatments more effective. This research contributes to the field of transportation safety by integrating traditional statistical methods with modern predictive analytics, providing a comprehensive, proactive approach to roadway safety decision-making.
- 언어주기
- English
- 일반주제명
- Statistics
- 일반주제명
- Transportation
- 일반주제명
- Information science
- 키워드
- Traffic volume
- 기타저자
- The University of Texas at Austin Civil Architectural and Environmental Engineering
- 기본자료저록
- Dissertations Abstracts International. 87-06A.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260311s2025 us eng d■001000017361235
■00520260311091547.5
■006m o d
■007cr|nu||||||||
■020 ▼a9798270232641
■040 ▼aMiAaPQD▼beng▼cMiAaPQD▼erda
■082 ▼a519.72
■1001 ▼aHasan, Manar▼eauthor.
■24510▼aCrash Modification Factors, Predictive Hotspot Identification, and Treatment Selection in Texas and Peer States ▼cManar Hasan
■260 ▼a[Sl]▼bThe University of Texas at Austin▼c2025
■264 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a1 electronic resource (283 pages)
■336 ▼atext▼btxt▼2rdacontent
■337 ▼acomputer▼bc▼2rdamedia
■338 ▼aonline resource▼bcr▼2rdacarrier
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-06, Section: A.
■500 ▼aAdvisors: Machemehl, Randy B. Committee members: Bhasin, Amit; Boyles, Stephen; Jiao, Junfeng.
■5021 ▼bPh.D.▼cThe University of Texas at Austin▼d2025.
■520 ▼aSafety is crucial in transportation, with huge human and property costs associated with every crash. Proactive data-driven decision-making is key in determining safety treatments that will have maximum positive impact. Traditional processes rely on historical crash data, generalized Crash Modification Factors (CMFs), and reactive interventions, often limiting their effectiveness. This dissertation addresses three challenges in safety decision-making: (1) assessing the reliability of currently used CMFs in estimating crash reductions, (2) identifying future high-risk crash locations, and (3) analytically performing treatment selection to maximize safety benefits. First, this dissertation evaluates the accuracy of CMFs currently used by transportation agencies by comparing expected crash reductions with observed outcomes. The findings reveal that some CMFs closely align with real-world crash data, while others, may overestimate or underestimate safety benefits. Also, CMFs can vary significantly for different crash severity levels, so jurisdictions that use one generalized value are inaccurately estimating safety treatment effects. Second, the dissertation performs predictive hotspot identification, leveraging Poisson and Random Forest models to forecast crash-prone locations. The models demonstrate good predictive performance, with traffic volume (ADT), traffic factors (peak hour factor, directional factor), speed limits, and surface width emerging as key factors influencing crash risk. The results suggest that integrating statistical and machine learning analyses into safety planning can aid proactive crash prevention, shifting from traditional reactive approaches. Finally, the dissertation proposes a quantitative, data-driven framework for treatment selection. Using propensity score matching (PSM) and predictive modeling with many models (Difference-in-Differences, Linear Regression, Random Forest, Gradient Boosting, Poisson, Negative Binomial, XGBoost, SVR, and Neural Network), the research evaluates treatment effectiveness across different roadway conditions and treatments. The findings highlight models that performed better than others (Poisson, Negative Binomial, Gradient Boosting, Random Forest, and XGBoost), and also some conditions which make certain treatments more effective. This research contributes to the field of transportation safety by integrating traditional statistical methods with modern predictive analytics, providing a comprehensive, proactive approach to roadway safety decision-making.
■546 ▼aEnglish
■590 ▼aSchool code: 0227
■650 4▼aStatistics
■650 4▼aTransportation
■650 4▼aInformation science
■653 ▼aCrash Modification Factors
■653 ▼aTraffic volume
■653 ▼aPropensity score matching
■653 ▼aPredictive hotspot identification
■7102 ▼aThe University of Texas at Austin▼bCivil, Architectural, and Environmental Engineering.▼edegree granting institution.
■7201 ▼aMachemehl, Randy B.▼edegree supervisor.
■7730 ▼tDissertations Abstracts International▼g87-06A.
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17361235▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
Preview
Export
ChatGPT Discussion
AI Recommended Related Books
Подробнее информация.
- Бронирование
- не существует
- моя папка
- Первый запрос зрения
- Non-Book Loan Application
- Nighttime Book Loan Application
Available after logging in.


