Method and device for determining interest degree of user in article, equipment and storage medium

Abstract

The invention provides a method and device for determining the interest degree of a user in an article, machine equipment and a computer readable storage medium. The method comprises the steps of obtaining classification behavior information representation of each classification behavior of a target user according to classification of the behaviors of the target user; obtaining an information representation of the candidate item; and determining the interest degree of the target user in the candidate article according to the classification behavior information representation of the classification behavior of the target user and the information representation of the candidate article. Through the embodiment of the invention, the determined interest degree can be closer to the real conditionof the user.

Description

technical field [0001] The present disclosure relates to the technical field of the Internet, and in particular to a method and device for determining a user's degree of interest in an item, a machine device, and a computer-readable storage medium. Background technique [0002] In Internet products, the recommendation system is widely used. It is generally based on big data and algorithms to determine or predict user preferences / interests, and recommend items that meet user preferences / interests as much as possible to improve the success rate of recommendations. Common recommendation methods can be divided into three ways: content-based recommendation, collaborative filtering-based recommendation, and cross-mixed recommendation. [0003] For example, in the field of news recommendation, the existing personalized intelligent recommendation methods are as follows: (1) analyze user logs, obtain user interest tags, and recommend news products of interest to users through tags; (...

AI Technical Summary

Problems solved by technology

[0005] The existing technology (1) recommends news products to users by updating tags, which has the advantages of simplicity and high efficiency, but the personalization effect is poor, and the tag definition is broad, which cannot accurately and fully reflect the inherent preferences of users for news, and is affected by Noise effect is obvious

[0006] The existing technology (2) is conducive to providing convincing recommendation explanations, but if there are many users, the cost of calculating the similarity matrix is ​​very high, and it faces the problem of data sparsity

[0007] In the existing technology (3), the recommendation is intuitive and does not require domain knowledge, but the recommendation result directly depends on the selection of features, and most of them only use the user's click on the product as the modeling standard

Images