Link Prediction Method Based on Common Neighbor Resource Allocation and Naive Bayes

Abstract

The invention discloses a link prediction method based on common adjacent node resource allocation and naive Bayesian. By establishing a network model G, any two unconnected nodes x and node y in the network and the common link between node x and node y are selected. Neighboring nodes, calculate the distribution value between node x and node y under the action of the common neighbor node; secondly, use the naive Bayesian method to obtain the connection attribute function of the common neighbor node in step S1, and use the connection attribute function to distinguish the common neighbor The role difference of the nodes; finally, combined with the distribution value between the node pairs to be predicted and the connection attribute function of the common neighbor nodes, the final similarity value of any two unconnected node pairs in the network is calculated, and the prediction is treated according to the final similarity value Node pairs for network link prediction. The invention uses the Naive Bayesian method to supplement the attribute difference between different nodes, so that the link prediction accuracy can be effectively improved.

Description

technical field [0001] The invention relates to the technical field of network science and technology and network link prediction, in particular to a link prediction method based on common adjacent node resource allocation and naive Bayesian. Background technique [0002] The rapid development of cities has formed a variety of complex networks around us, such as social relationship networks, economic networks, transportation networks, power networks, etc. The increasingly networked society requires us to understand all kinds of artificial and natural complexities. A better understanding of network behavior. Network science provides us with a new perspective and a new method for studying complex networks. With the increasing development and popularization of network science, people's understanding of complex networks is getting deeper and clearer. Link prediction is an important branch of network science. It mainly studies two aspects: on the one hand, it predicts some links...

AI Technical Summary

Problems solved by technology

However, the information investigated by these algorithms is too limited, the extraction and mining of the network structure are not enough, the problem of resource allocation of common neighbor nodes is not fully explored, the important role of common neighbor nodes in two unconnected nodes is not deeply explored, and there is no Considering the different effects of different attributes of the nodes themselves on the generation of links, it is impossible to effectively distinguish the different effects of common adjacent nodes on unconnected nodes on the connection.

Traditional neighbor-based methods do not have high prediction accuracy in actual networks

Images