Commodity recommendation method based on selective neighborhood information

Abstract

The invention discloses a commodity recommendation method based on selective neighborhood information. The method comprises the steps of obtaining interest preferences of a target user; obtaining attribute characteristics of the commodity; obtaining all users having historical score records with the commodity, taking all users as neighborhood information of a target user under the commodity, and determining interests and preferences of the users; determining the similarity between the target user and the interest preference of each user in the neighborhood information; fusing the interest preferences of the neighborhood information to obtain neighborhood feature information of the target user; determining the availability of the neighborhood feature information; obtaining target user feature information based on neighborhood information; obtaining a prediction score of the target user for the commodity; and selecting a plurality of commodities with the maximum prediction score to recommend to the target user. According to the method, the neighborhood information serves as auxiliary information, similar neighborhood information is automatically filtered out, the consistency of interests and preferences between the user and the neighborhood information of the user is captured, the data sparsity problem can be effectively relieved, and the recommendation precision is remarkably improved.

Description

Technical field [0001] The present invention relates to the technical field of commodity recommendation systems, in particular to a commodity recommendation method based on selective neighborhood information. Background technique [0002] Personalized recommendation technology mainly analyzes the behaviors of different users, guesses user interests, and actively recommends resources to users, thereby alleviating the conflict between the explosion of Internet information and users' quick access to information, and also making up for the weak ability of general search engines to personalize feedback results. Disadvantages. [0003] With the continuous development of e-commerce and the continuous expansion of scale, the number and types of products are also rapidly increasing, which makes customers spend a lot of time in selecting the products they need, and they will buy them in a relatively short time. Of products have become the development direction of online shopping. Recommend...

AI Technical Summary

Problems solved by technology

[0006] Although the existing recommendation methods can improve the recommendation effect by using the neighborhood information as auxiliary information; however, the existing recommendation methods introduce the neighborhood information into the recommendation process indiscriminately, ignoring the impact of the neighborhood information on the user. There are individual differences in the usability of the user. Because some users' interest preferences do not depend on the user's neighborhood information, the neighborhood information is introduced without distinction. On the contrary, it introduces interference information into the recommendation process, affects the recommendation performance, and makes the recommendation In addition, when the existing recommendation method introduces neighborhood information, it does not explicitly model the semantic similarity between the user and the user's neighborhood, resulting in the inability to efficiently capture the user's relationship with the user. The similarity between the neighbors, the recommendation effect is not good

Images