A method for constructing user portraits of microblog heterogeneous information

Abstract

The invention relates to a user portrait construction method for microblog heterogeneous information, which belongs to the field of machine learning and deep learning. The method includes: 1) collecting suitable text and image data, and training image and text classification networks respectively; 2) using crawler technology to crawl the text and image information of Weibo, and using the constructed classification network to obtain labels, namely user- 3) use fuzzy clustering on user-topic to discover user communities; 4) use TF-IDF to discover keywords on user-classified microblogs, get topic-item matrix, and use knowledge Graph thinking calculates associated words; 5) Calculate the user similarity to the community to which the user belongs, select users with high similarity to repeat the method of step 4), and select keywords according to the word frequency threshold as content supplementary recommendations. The invention can recommend users with high similarity for users.

Description

technical field [0001] The invention belongs to the field of machine learning and deep learning, and relates to a method for constructing user portraits of microblog heterogeneous information. Background technique [0002] In the era of big data, with the rapid increase in the amount of data, Weibo users are facing problems brought about by the information explosion. It is particularly important to understand user characteristics by establishing user portraits for community discovery and community recommendation. There is a large amount of heterogeneous information in Weibo data, such as pictures, videos and texts. This information is effectively extracted through deep learning methods, and used as user characteristic information to form user portraits, which can facilitate community discovery and product recommendation for users. [0003] The existing microblog user portrait creation method mainly extracts user keywords through TF-IDF and LDA methods, which cannot briefly d...

AI Technical Summary

Problems solved by technology

[0006] In view of this, the purpose of the present invention is to provide a method for constructing user portraits of heterogeneous microblog information, which effectively utilizes heterogeneous microblog content to construct accurate interest portraits of microblog users, and solves the problem of inconsistencies in traditional modeling methods with fine-grained themes The problem

Images