Network recovery method, device, server and storage medium

Abstract

The invention discloses a network recovery method and device, a server and a storage medium, and belongs to the technical field of Internet. According to the invention, fewer network samples are used;compared with the prior art, the method has the advantages that the approximate low-rank characteristic of the network is utilized, the optimization problem is established with the minimization of the nuclear norm as the target, the obtained optimal solution is perfected through subsequent adjustment steps, finally, the network recovery purpose is achieved, and due to the fact that the characteristics of the network tend to be similar, the method is higher in practicability and can be applied to various different use scenes. Furthermore, the optimization problem of the minimized nuclear normcan greatly save computing resources and human input, improve the analysis speed and bring possibility to more real-time application.

Description

technical field [0001] The invention relates to the technical field of the Internet, in particular to a network recovery method, device, server and storage medium. Background technique [0002] With the development of Internet technology, data of various network structures may be encountered, such as social networks, information dissemination networks, sensor networks, communication networks, gene / protein structure networks, etc. In many cases, due to the limitations of data acquisition methods or the particularity of network data, only part of the network may be obtained, that is, network data with missing nodes or edges, which makes it difficult to obtain information and apply information. Therefore, there is a need for a method that can infer the missing part of the network information based on the part of the network. This method is called the network restoration method. [0003] The current network recovery method is generally: to solve the network recovery problem bas...

AI Technical Summary

Problems solved by technology

[0005] When using the above similarity algorithm to solve the problem of network recovery, taking the recommendation system as an example, different types of recommendation systems, such as text recommendation and image recommendation, have different similarity metrics, so a related network recovery method can only To solve the recommendation problem of a single aspect, for different application scenarios, different corresponding network recovery methods are required, resulting in limitations in network recovery methods

Images