主催: Eastern Asia Society for Transportation Studies
p. 355
Road accident is one of the major causes of death and economic losses in Thailand. One of the steps to address the accident problem is to understand with proper depth how the accident can occur and identify hazardous locations in order to countervail accidents and reduce the severity of accidents on the highways. For that, a traditional practice in the developed world is to develop an accident prediction model. However, this is always associated with the availability of data and remains as a tool considered by many researchers to be suitable for the developed world only. This paper investigates the possibilities to develop a simple accident prediction model for Thai highways constrained by substantial lack of data and inspect its supremacy as compared with well known frequency based approach to identify hazardous locations. Hence, Poison regression models by maximum likelihood method were developed incorporating available geometric data, month, and exposure.