Research on the prediction of traffic accident by linear regression

Abstract

Traffic accident is getting increasingly serious. Although previous researchers use a variety of methods to predict the traffic accident, there are numerous demerits that need to be improved. This article demonstrates 12 variables that impact the traffic accident with 679 samples of accidents in UK from 2012 to 2014. This paper first analyses the relevance between dependent and independent variables, and also two independent variables to show the correlation between each factor. By using the multiple linear regression, it is concluded that although some independent variables do not have relationship with the dependent variable urban or rural area, Accident Severity, Number of Casualties, Road Type, Speed limit, Junction Control show significant relationship with the dependent variable. The paper also considers the 95% confidence interval in order to compare the effective density of data. Overall, the prediction of traffic accident is based on a number of factors and a sizable sample of accidents to summarize the impact that traffic accidents bring.