Modelling Road Accident using Poisson regression Models
The objective of this research is to evaluate the safety of multi-lane urban roads in India. In this paper, a generalised linear modelling technique is applied for the analysis of the Indian Highway's road accident. The features of road, speed, and traffic data are analysed in Surat on four-lane urban roads. A novel approach to the model of accident prediction for an urban highway is being proposed to include daily average travel (ADT) and average spot speed (AS). The model was developed as a reliant variable and significant variables such as chain width, intersection no, ADT, AS, as separate variables for accidents per kilometre.. The results of the model provide a better assessment of accidents on a multilateral urban road. Because road accidents are different, statistical models do not adequately capture the characteristics of each section. As a result, the results of Poisson regression were the opposite of these variables. There was also no statistically significant type of traffic control. Significant statistically at level 0.05. Accident locations were assessed by correlating the severity of the accident with different attributes. This investigation will contribute to improving urban road safety.