Multi-label propagation method and system based on implicit association

Abstract

The invention relates to a multi-label propagation method and system based on implicit association. The method comprises the following steps of constructing a local network structure diagram for a given network, and calculating a probability transfer matrix, wherein the label information of part of nodes is known; mining implicit association information among the plurality of label based on a label co-occurrence method; generating node-label matrix and initializing label information of the unknown node; updating the label of each unknown node according to the probability transfer matrix and the implicit association information; calculating an updating stop condition based on the node-label matrix; repeatedly executing the updating steps until an updating stop condition is met or a given number of iterations is reached; generating label information of nodes with unknown label information in the network according to the node-label matrix . According to the method, the association relationship between the labels can be better mined, and the convergence speed of the method is accelerated, so that the label information of the large-scale network user is analyzed more accurately and comprehensively.

Description

technical field [0001] The invention belongs to the technical field of social network applications, and in particular relates to an implicit association-based multi-label propagation method and system. Background technique [0002] In the real world, there are common connections and mutual dependence between users. In many existing large-scale networks, a large number of node label information is missing, and people can only estimate through a small amount of node label information. With the development of computer-related technologies, more and more methods are used to model complex large-scale networks, explore the relationship and potential laws between different nodes, and better predict the unknown label information of nodes. The emergence of natural language processing (NLP) technology provides a favorable tool for mining the explicit relationship between tags. [0003] So far, many multi-label propagation algorithms have been proposed. The label propagation algorith...

AI Technical Summary

Problems solved by technology

The advantage of this algorithm is that the calculation process is simple and the method is fast, but the disadvantage is that the stability of the algorithm is poor, and the results are very different each time. At the same time, the relationship between labels in the real world is intricate, but this method does not consider the relationship between labels. relationship, therefore, the actual application value is not high

With the development of NLP technology, people use NLP tools to mine the explicit information between tags, which improves the accuracy of the results. However, in the real world, there are still unknown implicit associations between tags. Therefore, the existing multi-label There is still a lot of room for improvement in the propagation method

Images