The prone points extraction method of road traffic accident based on the spatiotemporal kernel density of the network

In order to overcome the inaccuracy and low accuracy of the evaluation results of the accident risk degree in the prone points extraction process of the road traffic accident, this paper proposes a new prone points extraction method of the road traffic accident based on the spatiotemporal kernel density of the network. This method builds the spatiotemporal model of the road traffic network, and simplifies the model according to the expression of the model feature data. The real probability density function of random variable is obtained by introducing the spatiotemporal kernel density estimation method of network, so as to evaluate the risk degree of accident prone point and complete the calculation of the risk degree limit value of accident prone point. According to the results of the calculation, the prone points extraction system of road traffic accident is established, and the prone points extraction of road traffic accident is realized based on the clustering results of accident risk degree. The experimental results show that compared with the three traditional prone point extraction methods of road traffic accident, the new prone point extraction method of road traffic accident greatly improves the accuracy of extraction, and the convergence effect of extraction error is better. The experimental results fully show that the proposed method has better application effect.