THE EFFECT OF METEOROLOGICAL FACTORS ON ROAD TRAFFIC INJURIES IN BEIJING

The goal of this study is to establish the relationship between RTI and the meteorological factors, and make a precise prediction of the occurrence of RTI in China The statistical data was collected from four tertiary hospitals in Beijing from 2008 to 2012. The association was analyzed using a backward propagation-artificial neural network model (BP-ANN) regression model (built by Matlab) which has three layers (including one hidden layer). Based on the SPSS 20.0 platform, regression analysis was used to evaluate the effect of meteorological factors on RTI. The results show that the most significant factors are atmospheric pressure, temperature, precipitation and sunshine duration, whereas wind speed is less significant. The combination of long sunshine duration, high temperature, low pressure and high humidity is the high-risk condition that leads to RTI. The coefficient (r = 0.7199) obtained by the PB neural network is much higher than the coefficient (r = 0.427) obtained by the Stepwise Regression Model. Meteorological factors have a certain effect on traffic injury severity. And the BP- ANN model is a quite precise prediction model for RTI, and this research can provide technical support for the improvement of forecasting and early warning of RTI, in further.