D.X. Liu

Pages: 57-66

Abstract
There are many problems in traffic accident data statistics, such as large statistical error, high data noise and long time-consuming. The spatial data statistical model of urban road traffic accident is designed. Firstly, determine the distribution state of traffic accident spatial data, extract the traffic accident spatial data under the uniform distribution state with the help of Bayesian network, and transform the randomly distributed accident spatial data into linear data; Then, the decision tree is used to calculate the information entropy of noisy traffic accident spatial data, determine the proportion of noisy data, and complete the preprocessing; Finally, the objective function is used to optimize the data membership, construct the spatial data statistical model of urban road traffic accidents, and complete the design of the spatial data statistical model . The experimental results show that the statistical data error of the design model is less than 2%.
Keywords: accident spatial data; statistical model; distribution status; bayesian network; noisy data; correction function