Four-dimensional tensor decomposition recommendation method fusing comment text and feature weighting

Abstract

The invention discloses a comment text and feature weighting fused four-dimensional tensor decomposition recommendation method. The method comprises the following steps: S1, carrying out data acquisition and division; s2, processing the comment text; s3, constructing a tensor; s4, carrying out weighting in combination with labels and score data; s5, carrying out tensor decomposition and recommendation generation. According to the method, comment texts are integrated into a tensor decomposition model. The method includes collecting users, user comment texts, items and item comment texts by using the user item comment data set, and vectorizing the comment texts; respectively obtaining a feature vector of a user and a feature vector of an item, and then constructing a four-dimensional tensor {the user, the user feature, the item and the item feature}; and finally, fully mining the potential relation between tensor entities by applying high-order singular value decomposition, and generating recommendation according to the processing result, thereby achieving the purpose of improving the efficiency of the recommendation system.

Description

technical field [0001] The invention belongs to the technical field of data mining and natural language processing, and in particular relates to a four-dimensional tensor decomposition recommendation method which combines comment text and feature weighting. Background technique [0002] With the rapid development of information technology, people gradually enter the era of information overload, how to find the information they need among the overloaded information is particularly important. The function of the recommendation system is to analyze the behavior activities and needs of the target users, and then recommend items that they may like. Therefore, improving the accuracy of the recommendation system can bring huge economic benefits to the enterprise and allow users to have a better experience. [0003] In recent years, recommendation systems have been favored by more and more industries, especially the development of personalized recommendation systems, which has play...

AI Technical Summary

Problems solved by technology

When the recommendation model directly uses the scoring data to participate in the recommendation, the model will be ineffective due to the sparsity of the scoring data. One way to alleviate the data sparsity is to use the comment data to make up for the lack of data by supplementing more information.

Images