Intelligent analysis system for organizational relationship based on knowledge graph and deep learning

CN122241178APending Publication Date: 2026-06-19云南省标准化研究院

Patent Information

Authority / Receiving Office: CN · China
Patent Type: Applications(China)
Current Assignee / Owner: 云南省标准化研究院
Filing Date: 2026-03-16
Publication Date: 2026-06-19

AI Technical Summary

Technical Problem

Existing intelligent analytics technologies struggle to accurately predict business rules when processing massive and dispersed organizational data. Deep learning models have weak logical coherence, making it difficult to accurately predict business rules, while knowledge graph models have poor generalization ability and struggle to extract deep patterns from the data, resulting in inaccurate analysis of organizational relationships.

Method used

A closed-loop feedback iteration mechanism based on knowledge graphs and deep learning is adopted. A standardized dataset is constructed through a multi-dimensional data acquisition module, organizational relationships are analyzed using a knowledge graph construction module, association patterns are mined using a deep learning module, and logical verification and parameter updates are performed through a logical verification feedback module to achieve high-precision analysis of organizational relationships.

Benefits of technology

It improves the accuracy and interpretability of organizational relationship analysis, and can accurately output hierarchical relationship tree diagrams and investment relationship network diagrams, discover hidden business relationships and equity structures, and support intelligent monitoring and risk warning.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure CN122241178A_ABST

Patent Text Reader

Abstract

This invention relates to the field of big data analytics and discloses an intelligent organizational relationship analysis system based on knowledge graphs and deep learning. The system includes a multi-dimensional data acquisition module, a knowledge graph construction module, a deep learning module, and a logic verification feedback module. First, the system cleans and encodes multi-source, dispersed data to construct a standardized organizational dataset, and then builds an initial knowledge graph based on this dataset. Next, it uses a deep learning network to mine data features and outputs relationship prediction results. Subsequently, the prediction results are input into the knowledge graph for logical verification. If rigid logic is violated, a logic verification error is generated. This error is fused with the prediction error to form a joint loss feedback signal, which is backpropagated to update the network parameters until convergence. This invention overcomes the weakness of deep learning's logical integrity, outputting high-precision hierarchical relationship tree diagrams and investment network diagrams, achieving in-depth analysis and risk warning.

Need to check novelty before this filing date? Find Prior Art