Intelligent Network Boundary Division Based on K-Means and DBSCAN Clustering Features

The modeling and control of converged networks has received extensive attention in recent years. The division of road network boundary is one of the most important steps in urban traffic control. The traditional division method is mainly based on administrative region and several division principles. Based on the idea of cluster analysis and the spatial topology characteristics of road segments in the road network, we analyzed the effects and the applicability of K-means algorithm and DBSCAN algorithm on road network boundary division. Converging road network traffic density and traffic flow and compare the preliminary segmentation, to obtain the final road network segmentation results. We select two road network division indicators to evaluate the effect of two division methods on road network boundary division. The results show that the classification coefficients of the two types of algorithms are 0.0042, 0.0056. The K-means algorithm is used to determine the boundary of the road network based to the density. The variance indicators before and after the adjustment are 0.7476 and 0.7442 respectively, so the overall variance is reduced. Therefore, the road network segmentation after the boundary adjustment is better. Compared with the DBSCAN partitioning method, the K-means algorithm refine and obtain better road network segmentation results.