Rough set approach for accident chains exploration
This paper presents a novel non-parametric methodology - rough set theory - for accident occurrence exploration. The rough set theory allows researchers to analyze accidents in multiple dimensions and to model accident occurrence as factor chains. Factor chains are composed of driver characteristics, trip characteristics, driver behavior and environment factors that imply typical accident occurrence. A real-world database (2003 Taiwan single auto-vehicle accidents) is used as an example to demonstrate the proposed approach. The results show that although most accident patterns are unique, some accident patterns are significant and worth noting. Student drivers who are young and less experienced exhibit a relatively high possibility of being involved in off-road accidents on roads with a speed limit between 51 and 79 km/h under normal driving circumstances. Notably, for bump-into-facility accidents, wet surface is a distinctive environmental factor.
- Record URL:
-
Availability:
- Find a library where document is available. Order URL: http://worldcat.org/issn/00014575
-
Supplemental Notes:
- Abstract reprinted with permission from Elsevier
-
Authors:
- Wong, Jinn-Tsai
- Chung, Yi-Shih
- Publication Date: 2007-5
Language
- English
Media Info
- Media Type: Print
- Features: Figures; References; Tables;
- Pagination: pp 629-637
-
Serial:
- Accident Analysis & Prevention
- Volume: 39
- Issue Number: 3
- Publisher: Elsevier
- ISSN: 0001-4575
- Serial URL: http://www.sciencedirect.com/science/journal/00014575
Subject/Index Terms
- TRT Terms: Crash causes; Crash characteristics; Drivers; Nonparametric analysis; Traffic crashes
- Uncontrolled Terms: Rough set theory
- Geographic Terms: Taiwan
- Subject Areas: Data and Information Technology; Design; Highways; Safety and Human Factors; Vehicles and Equipment; I81: Accident Statistics; I82: Accidents and Transport Infrastructure; I83: Accidents and the Human Factor; I91: Vehicle Design and Safety;
Filing Info
- Accession Number: 01051765
- Record Type: Publication
- Files: TRIS
- Created Date: Jun 19 2007 8:28AM