Convolutional neural network-based user attribute inference method and apparatus

Abstract

The invention relates to a convolutional neural network-based user attribute inference method and apparatus. The method comprises the steps of establishing a self-centered network according to attributes of user nodes and a friend relationship; and extracting attribute information of the user nodes in the self-centered network and hidden information comprised in the friend relationship by adoptinga convolutional neural network, and inferring missing attributes of a user by utilizing the hidden information. For a social network which cannot directly obtain the friend relationship or has relatively high obtaining difficulty, the missing attributes are subjected to classification prediction by only utilizing the attribute information of the user through adopting the neural network. The limitation of manual definition of a similarity function can be well avoided; and relationships among different attributes and different attribute dimensions can be better showed through convolution operation of a convolution kernel, so that the missing attributes of the user can be inferred efficiently and accurately.

Description

technical field [0001] The invention belongs to the technical field of user missing attribute inference in social networks, especially the inference of multi-valued attributes, and specifically relates to a method and device for inferring user attributes based on a convolutional neural network, which has high accuracy. Background technique [0002] With the development of Internet technology, online social products, such as Weibo, Zhihu and Facebook, have become a necessity in users' daily life. While using these products, users generate a large amount of information, including user attribute information, post content, and friend relationships. This information provides data support for companies and scientific researchers to accurately describe user portraits. At the same time, in order to protect the privacy of users, online social products provide users with fine-grained privacy settings, which makes it difficult to directly obtain user attribute information. According t...

AI Technical Summary

Problems solved by technology

The limitation of the classification-based method is that the calculation of the similarity needs to be comprehensive and accurate, and the accuracy of the classification model must be high, and the construction of data features must be comprehensive. However, although the accuracy of the machine learning model in the current method is high, most of them are aimed The problem of binary classification, and in the social network, it is faced with the problem of not being able to obtain a comprehensive feature vector of multi-dimensional attributes to construct users, which leads to poor prediction results, especially the prediction of multi-valued attributes.

However, the attribute prediction algorithm based on label propagation needs to spend a lot of time to calculate the adjacency matrix of the graph, and the algorithm itself treats the importance of friends of the marked nodes equally, which itself does not conform to the characteristics of social networks, so in real data The effect is also unsatisfactory

Images