Link prediction method and device in urban road network

Abstract

The invention belongs to the technical field of urban traffic, and particularly relates to a link prediction method and device in an urban road network. The method comprises the steps of collecting anurban road network data set, and calculating an attribute value of each urban place; dividing each attribute value into class clusters by utilizing a cloud model to obtain a dichotomy result about whether link connection is established between the attribute and the city node or not, and calculating the determinacy of each attribute for the dichotomy label; calculating a frequent binomial set of the two associated attributes according to a binary classification result to construct a two-layer functional network model structure; training the two-layer functional network model, and minimizing atarget loss function by adopting a least square method; inputting urban road network data to be tested into the trained two-layer functional network model, and outputting a binary classification labelfor judging whether a road is generated between two urban places in the urban road network; according to the invention, the high dependence of a complex network structure is overcome, and the link prediction effect is optimized, so that the prediction accuracy is improved.

Description

technical field [0001] The invention belongs to the technical field of urban traffic, and in particular relates to a link prediction method and device in an urban road network. Background technique [0002] With the development of urbanization, the structure of urban traffic road network will have obvious differences due to different traffic modes, natural conditions, population distribution, evolution stage and development level. The development of economy and technology changes the road network structure all the time. The evolution of the road network (hereinafter referred to as "road network") will become more and more complex. The complex evolution of the road network has increased the difficulty of decision-making in urban road planning. If the road planning is unreasonable, it will lead to an unreasonable allocation of road network resources, and may even hinder the development of urbanization. [0003] Since the 20th century, with the rise of research on "small world...

AI Technical Summary

Problems solved by technology

[0004] At present, it can be roughly divided into two methods of link prediction: the first is the link prediction technology based on the traditional method, which mainly uses the similarity between nodes to predict, and uses the connection between nodes to measure their similarity. This is used to learn and predict the connection; the second method is the link prediction based on deep learning, which mainly uses the graph neural network to predict the urban road network, and describes the embedding of nodes by performing convolution operations on the embedding of nodes , for example, Chinese patent CN108108854A proposes a method, system and storage medium for urban road network link prediction. This patent constructs the adjacency matrix of the road network, obtains the Katz similarity matrix according to the adjacency matrix, and converts the Katz similarity After the matrix is ​​normalized, use a multi-layer nonlinear autoencoder to perform network representation learning on it to obtain a network representation vector, decode and reconstruct the adjacency matrix according to the network representation vector, and perform road network linking according to the reconstructed adjacency matrix. prediction; however, the deep learning model used in the above patents has uncontrollable problems, so it is necessary to comprehensively consider the model effect and model controllability to achieve effective link prediction

Images