User portrait prediction method based on multi-source transboundary data fusion

Abstract

The invention discloses a user portrait prediction method based on multi-source cross-boundary data fusion, and aims to solve the problem of inaccurate user feature prediction caused by item feature sparsity, high-order structured feature deficiency and user behavior sequence feature deficiency in the prior art. Based on the e-commerce data generated by the user, commodity features are expanded by using the knowledge graph, historical purchase records of the user are fully mined by using the graph convolutional network, and potential purchase features of the user are predicted by using the recurrent neural network, so that the accuracy of user portrait prediction is effectively improved. The method has the advantages that the problem of commodity feature sparsity is solved through the knowledge graph, the problem of high-order structured feature deletion is solved through the graph convolutional neural network, the problem of user behavior sequence feature deletion is solved through the recurrent neural network, and a good foundation is laid for performance improvement of a recommendation system.

Description

technical field [0001] The invention relates to a user portrait prediction method based on multi-source cross-border data fusion based on historical order sequences and content information of user shopping. Background technique [0002] With the development of Internet technology and smart devices, various mobile applications have appeared and penetrated into people's lives, and the information generated has also exploded. This makes it difficult for people to efficiently obtain the information they need, and it is difficult for companies to accurately push products or information to users. The basis of the recommendation system is user portraits. Efficient construction of user portraits will help enterprises achieve refined marketing and precise recommendations. [0003] Online shopping has become an extremely common thing in today's life. While enjoying the convenience brought by online shopping, many users also directly or indirectly provide personal information to shopp...

AI Technical Summary

Problems solved by technology

The hidden features of the above items often have sparsity problems in network platforms

In addition, most of the above did not mine the association between users and items, and most of them regard user feature prediction as a classification task, and each feature of the user is relatively independent, resulting in a certain degree of loss of the associated features between users and items , cannot effectively learn a user representation vector for user feature prediction

Images