Forecast Study for Road Traffic Accident based on Neural Network

摘要

As the research and practice proved, it is feasible to predict the changes and the developments of the traffic accidents scientifically by recognizing and using objective law of the development of the traffic accident. Artificial neural network has the advantages of nonlinear approximation, fuzzy reasoning, large-scale parallel processing, self-learn and good fault-tolerance. After analyzing and determining the road accident factors, BP neural network model is built to predict the road traffic accidents. The forecast results indicate that the structure and parameters of the built model are practical for road traffic accidents of China.