A Method for Determining the Number of Restored Edges in Link Prediction

Abstract

The invention belongs to the fields of network reconfiguration and link prediction and particularly relate to a method for determining the quantity of link prediction restitution edges. The method comprises the steps that based on the similarity of link prediction and edge clustering, the quantity of the restitution edges in the link prediction process is determined by utilizing segmentation density indexes in edge clustering; connecting probabilities of all node pairs of which the edges are not connected are calculated based on a link prediction technology, and sequencing is performed according to the connecting probabilities; according to the sequence, all edges are sequentially added in a network in a tentative mode, and edge clustering analysis is performed on the network again; in the edge clustering process, the segmentation density indexes are calculated; through the change of the maximum segmentation density before and after a certain edge is added, whether the edge is added or not is determined. According to the method, the problem existing in determination of the quantity of the restitution edges after link prediction is solved; besides, the method is simple in principle, clear in process, easy to implement and capable of providing support in decision for a decider and improving the decision efficiency.

Description

technical field [0001] The invention belongs to the field of network reconfiguration and link prediction, and in particular relates to a method for determining the number of link prediction restoration edges. Background technique [0002] What link prediction needs to solve is the problem of restoring and predicting missing edges in complex networks. The existing link prediction methods, due to the consideration of computational speed and computational complexity, use similarity-based chains in practical applications. road forecast. Existing link prediction techniques cannot decide how many edges should be restored. Through the existing method, the decision maker can only calculate the connection probability and the order of the connection probability of the unconnected edge (no edge is directly connected in the observation network), and does not know the number of edges that should be restored. It is a problem that must be solved in the process. [0003] Several existing...

AI Technical Summary

Problems solved by technology

Since the existing link prediction method only gives the connection probability of unconnected node pairs in the network, there is no method to determine how many edges to restore. For decision makers, a method that can guide the number of restored edges is of great importance. the meaning of

Images