User tag identification method and device based on artificial intelligence, equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence, and provides a user tag identification method and device based on artificial intelligence, equipment and a medium. The method comprises the following steps: acquiring related information (such as behavior information of an online medical platform) of a user group and constructing a topological network of the user group; inputting the topological network into a graph convolutional neural network, calculating a first-order similar matrix, a second-order similar matrix and a Dice similarity matrix of the topological network, and calculating a target similar matrix of the topological network; calculating a low-dimensional matrix of the topological network based on the target similar matrix; taking a first type of nodes corresponding to the low-dimensional matrix as a sample set, and training a graph convolutional neural network to obtain a target classification model; and identifying second type nodes of the low-dimensional matrix based on the target classification model to obtain target labels of users corresponding to the second type nodes. According to the invention, the accuracy of node label identification can be improved. The invention further relates to the technical field of block chains, and the target label can be stored in a node of a block chain.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to an artificial intelligence-based user label identification method, device, equipment and storage medium. Background technique [0002] At present, the existing node classification methods are mainly methods based on matrix decomposition or random walk. Matrix factorization-based methods are not suitable for node classification tasks in large-scale networks because frequent matrix operations are involved in the process of matrix factorization, resulting in high system resources and time overhead. Although the method based on random walk can overcome this, the method based on random walk is mainly oriented to the topology of the network and ignores the attribute characteristics of nodes (for example, the behavior attribute information of users on the online medical platform, etc.), resulting in information Incomplete utilization affects the accuracy of node...

AI Technical Summary

Problems solved by technology

[0003] In view of the above, the present invention provides an artificial intelligence-based user label identification method, device, equ...

Images