| Summary |
Ying-Chun Liu, 2008.06
Department of Transportation Technology and Management National Chiao Tung University
Logistic regression and rough set theory are applied to analyze the relation between types and factors of fatal traffic accidents involving college students in the study. A sample of over 200 accident cases is included. The data are classified into freshman and non-freshman accident according to the grade of the students. Chi-squared test of independence is used first to identify key factors affecting accident types, and then logistic regression model is constructed to determine the odds of being involved in a single vehicle accident. Besides, rough set theory is utilized to propose accident casual chains which imply the logical links between factors and accident types. The results show that students’ vehicle and mental conditions and road types are three significant factors which have great influences on accident types. Freshmen are more likely to be involved in a single vehicle accident when driving a car, while non-freshmen are less likely to be involved when in good mental condition or at an intersection. Moreover, the frequently-occurred single vehicle accident patterns are found: freshmen in good mental condition drive/ride on a straight road in the suburbs; non-freshmen with motorcycle licenses drive/ride at night when exhaustively tired. |
