Recommendation list generation methods and devices

CN113934932BActive Publication Date: 2026-04-03BEIJING BAIDU NETCOM SCI & TECH CO LTD
View PDF -1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-10-15
Publication Date
2026-04-03

AI Technical Summary

Technical Problem

In existing recommendation methods, the resources retrieved may not be entirely based on user interests, but rather on meeting business needs, such as distributing trending events, resulting in insufficient accuracy and diversity in the recommendation list.

Method used

By obtaining the first set of content recalled by the recall model and the second set of content that meets the preset recall conditions, sorting and feature extraction are performed on them respectively to generate a recommendation list, ensuring that the ranking of content that users are interested in is not affected, while also meeting the insertion of content that meets business needs.

Benefits of technology

It improves the comprehensiveness and accuracy of the recommendation list, meets diverse needs, ensures that the order of user interest content remains unchanged, and reasonably inserts content based on business needs, thereby enhancing the diversity of the recommendation list and the system's ability to meet multiple objectives.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113934932B_ABST
    Figure CN113934932B_ABST
Patent Text Reader

Abstract

This disclosure provides a method and apparatus for generating recommendation lists, relating to the field of computer technology, specifically artificial intelligence and big data technology. The specific implementation scheme is as follows: First, the content to be recommended is obtained, including a first set of content recalled by a recall model and a second set of content that meets preset recall conditions. Then, the first set of content is sorted to obtain an initial list. Finally, the second set of content and the initial list are sorted to generate a recommendation list corresponding to the content to be recommended. This disclosure improves the comprehensiveness and accuracy of the recommendation list.
Need to check novelty before this filing date? Find Prior Art