Method and device for generating recommendation results

Abstract

Embodiments of the present invention provide a method and device for generating recommendation results. The method includes: obtaining item content information of at least one item and user scoring information of at least one user, where the user scoring information of a first user among the at least one user includes historical scoring of the at least one item by the first user; Using the item neural network and the user neural network to encode the item content information and the user scoring information respectively, and obtain the item latent vector of each item in the at least one item and the user of each user in the at least one user Hidden vector: determine the recommendation result for each user according to the item hidden vector and the user hidden vector. The method and device for generating recommendation results in the embodiments of the present invention can improve the accuracy of recommendation results, thereby improving user experience.

Description

technical field [0001] Embodiments of the present invention relate to the field of electronic services, and more specifically, to a method and device for generating recommendation results. Background technique [0002] In the past decade, the rapid development of the Internet and e-commerce platforms has brought a large amount of business data (user information, item information and scoring information). Through the analysis of these data, the items that users like can be predicted and recommended. Recommendation algorithms have been widely used in commercial systems, such as Amazon (Amazon), Netflix (Netflix), etc., and have brought a lot of revenue. [0003] The methods for generating recommendation results mainly include content-based recommendation, collaborative filtering-based recommendation and hybrid recommendation. Content-based recommendation mainly depends on the feature representation of the content, and generates a recommendation list in descending order accor...

AI Technical Summary

Problems solved by technology

However, in the CDL method, the user representation is still obtained through matrix decomposition, resulting in insufficient expressive ability of the user's hidden vector, and the final recommendation result is not accurate.

Images