Research On Forecasting Methods Of Road Traffic Accidents

The road traffic accidents seriously threaten people's life security and property security, so it becomes a public hazard in today's society. So far, the number of the traffic accident deaths in China has ranked first in the world over the years. Hundreds of thousands of people were killed or injured each year on average, with direct economic losses amounting to billions of dollars. The road traffic accidents forecasting is the important content of the research of traffic safety. The purpose of road accidents forecasting is to analyze the tendency of road traffic accidents in the future and take corresponding measures according to the traffic accidents forecasting results. To avoid the defects and deficiencies of the work in the future and ultimately achieve the purpose of reducing traffic accidents. Based on the analysis of existing forecasting methods of road traffic accidents, the principal component analysis and cluster analysis were introduced into the accidents forecasting in this paper. The main research contents and conclusions are as follows:1,A number of factors may affect the occurrence of road traffic accidents, and these factors may exist information overlap, which sometimes even obliterate the real traffic characteristics and the inherent laws, so the principal component analysis was introduced into the accidents forecasting. First, we can use principal component analysis to analyze the factors of influencing road traffic accidents, eliminate some overlapping information and achieve the purpose of dimensionality reduction. Then combining BP neural network to forecast road traffic accidents (BP Neural Network based on PCA), and finally to compare the results with non-principal component analysis of BP neural network. It can be concluded that BP neural network based on principal component analysis is obviously better than BP neural network on the precision.2,Different cities have different tendency of road traffic accident development in different stages. In order to better forecast road traffic accidents in different cities, the cluster analysis was introduced into road traffic accidents forecasting. Different cities will be divided into different categories using cluster analysis. Then using a variety of methods to forecast different types of cities, and finally to find suitable accident forecasting methods for each type of city.