Enterprise human resource performance evaluation method based on knowledge graph

By improving the Mamba architecture and neural logic reasoning network, and combining multi-source heterogeneous data with federated knowledge distillation methods, the problems of data silos and insufficient feedback in enterprise human resource performance evaluation are solved, realizing multi-dimensional and semantic enterprise-level performance evaluation and improving the accuracy and stability of the evaluation.

CN120672196BActive Publication Date: 2026-07-21JIANGSU VOCATION & TECHNICAL COLLEGE OF FINANCE & ECONOMICS
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
JIANGSU VOCATION & TECHNICAL COLLEGE OF FINANCE & ECONOMICS
Filing Date
2025-06-10
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing enterprise human resource performance evaluation methods suffer from problems such as single data dimensions, lack of semantic feature analysis, cross-departmental data silos, and insufficient feedback mechanisms, resulting in highly subjective evaluation results, low accuracy, and inability to achieve dynamic optimization.

Method used

By adopting an improved Mamba architecture and neural logic reasoning network, and through multi-source heterogeneous data acquisition, deep semantic feature extraction, dynamic logic rule learning, and federated knowledge distillation methods, cross-departmental collaborative analysis and feedback optimization are achieved. A multi-granularity logic reasoning knowledge graph library is constructed for enterprise-level global dynamic performance evaluation.

Benefits of technology

It enables multi-dimensional and semantic modeling of employee performance evaluation, improves the objectivity and accuracy of evaluation results, solves the problem of cross-departmental data silos, enhances the comprehensiveness and dynamism of evaluation, reduces evaluation errors, and improves the accuracy and stability of evaluation results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120672196B_ABST
    Figure CN120672196B_ABST
Patent Text Reader

Abstract

The application discloses a kind of enterprise human resource performance evaluation methods based on knowledge graph, comprising the following steps: S1, through the multi-source heterogeneous data interface acquisition staff daily work data, task collaboration data, process tracking data, enterprise knowledge base data and external market dynamic data, obtain initial multi-modal data;S2, to initial multi-modal data is washed and structured processing;S3, based on the improved Mamba architecture automatically extracts employee performance semantic features;S4, constructs multi-granularity logical reasoning knowledge graph library, and the weight of real-time learning and updating logical reasoning rule;S5, using dynamic logic rule set carries out neural logic reasoning, obtains employee performance grade score;S6, through federal knowledge distillation method realizes cross-department employee performance data collaborative analysis;S7, according to the deviation between evaluation result and actual performance feedback data optimizes performance evaluation result.The application realizes objective, dynamic and comprehensive performance evaluation.
Need to check novelty before this filing date? Find Prior Art