Enterprise multi-level knowledge tag recommendation method and equipment
By combining semantic extraction models and multi-level knowledge graphs, this approach addresses the issues of low efficiency, insufficient accuracy, and adaptability in enterprise knowledge tag management. It enables precise tag recommendation and knowledge asset management, and is applicable to hardware devices such as enterprise knowledge platforms and cloud server clusters, while also being suitable for multi-terminal knowledge management in large enterprise groups.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-11
- Publication Date
- 2026-03-31
AI Technical Summary
Existing enterprise knowledge tag management suffers from low tagging efficiency, insufficient accuracy, lack of systematic management, inability to adapt to the knowledge management needs of multi-level and multi-business lines in group enterprises, lack of scenario adaptability and real-time performance in intelligent recommendations, lack of multi-level compliance and collaborative control, absence of human-machine collaboration and iteration mechanisms, and insufficient AI friendliness of tags.
By parsing the business semantic information of knowledge documents through a semantic extraction model, matching candidate tag sets with a multi-level knowledge graph, and conducting comprehensive verification based on multi-level compliance verification rules, the relevance and hierarchical adaptability of tags and knowledge documents are realized. Based on user roles and business semantic information, multi-dimensional association reasoning is used to recommend compliant tags.
It enhances the relevance and practicality of tag recommendations, optimizes the efficiency of enterprise knowledge tag management and knowledge asset utilization, adapts to the multi-level management needs of group enterprises, and realizes improved efficiency of cross-level and cross-business line knowledge collaboration.
Smart Images

Figure CN121766434A_ABST