Loading...
Please wait, while we are loading the content...
Similar Documents
Individual Driver Risk Analysis Using Naturalistic Driving Data
| Content Provider | Semantic Scholar |
|---|---|
| Author | Guo, Feng Fang, Youjia |
| Copyright Year | 2011 |
| Abstract | Individual driver risk varies substantially and a small percentage of drivers often contribute to a disproportionate large number of safety events. Identifying factors associated with individual drivers and predicting high-risk drivers will benefit the development of driver education programs and safety counter-measures. The goal of this study is twofold 1) to assess risk factors associated individual drivers and 2) to predict high-risk drivers. The 100-Car Naturalistic Driving Study data was used for methodology development. A negative binomial regression analysis indicated that driver age, extroversion of the NEO 5 personality trait, and critical incidents, as a measure of aggressive driving behavior, had significant impacts on crash and near-crash risk. For the second objective, drivers were classified into three risk groups based on crash and near-crash rate using a k-mean cluster method. Approximately 6% of drivers were identified as high-risk and 18% of driver as high/moderate risk drivers. A logistic regression model was developed to predict high risk drivers as well as high/moderate risk drivers. The predictive models showed high predicting power with area under the curve value of 0.917 and 0.9351 for the receiver operating characteristic curves. This study concluded that age, personality, and driving behavior is closely related to individual driving risk and aggressive driving is a powerful predictor for high risky drivers. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://onlinepubs.trb.org/onlinepubs/conferences/2011/RSS/2/Guo,F.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
