A topic screening method, device, equipment and medium

By using large language models and topic knowledge graphs, topics on social media platforms are automatically filtered, solving the problem of low efficiency in manual topic filtering, achieving a highly efficient topic filtering process, reducing costs and improving efficiency.

CN122412618APending Publication Date: 2026-07-17ZHEJIANG BAORONG MEDIA TECH (ZHEJIANG) CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
ZHEJIANG BAORONG MEDIA TECH (ZHEJIANG) CO LTD
Filing Date
2026-06-16
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Existing topic selection methods rely on manual operation, resulting in high time and labor costs and low efficiency.

Method used

By acquiring topic selection instruction text, using a pre-trained large language model and topic knowledge graph, multiple related topics are automatically selected. Based on the basic indicators of related topics, value scores are calculated to determine candidate topics. Recommended topics are then selected through multi-dimensional quantitative evaluation values ​​and provided to editors.

Benefits of technology

It has automated the topic selection process, reducing time and manpower costs and improving topic selection efficiency. It can quickly select multiple topics with in-depth analysis and mining value from social media platforms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122412618A_ABST
    Figure CN122412618A_ABST
Patent Text Reader

Abstract

本发明涉及电数字数据处理技术领域，尤其涉及一种话题筛选方法、装置、设备及介质。方法包括：在获取到话题筛选指示文本之后，根据话题筛选指示文本和预先训练的大语言模型，从话题知识图谱中筛选出多个关联话题；确定各个关联话题的价值评分；根据价值评分在各个关联话题中筛选出多个候选话题，通过大语言模型，确定各个候选话题的话题参考信息；确定各个候选话题的多维量化评估数值；在各个候选话题中筛选出多个推荐话题，将推荐话题和话题参考信息提供给编辑用户。本发明实施例可以自动基于话题筛选指示文本、大语言模型以及话题知识图谱，快速从社交媒体平台展示的各个话题中筛选出具有深度分析和挖掘的价值的多个话题。
Need to check novelty before this filing date? Find Prior Art