Search

Design and Implementation of a Job-Skill Knowledge Graph and Talent Matching System Driven by Multi-Source Recruitment Data

Author(s)

Yinghuai Fu1, Shuai Li1,*, Zhen Hao2,*

Affiliation(s)

1School of Artificial Intelligence and Big Data, Henan University of Technology, Zhengzhou, China 2IFLYTEK Co., Ltd., Hefei, China *Corresponding Author

Abstract

Labor-market skill demand shifts faster than education cycles, while public job ads suffer lag, duplication, and noise. Addressing the challenge of multi-source cleaning with cross-validation and hallucination control for competency graphs, this paper presents Digital Talent Mapping with five auditable innovations on over twelve thousand real postings: (1)SHA256 content fingerprints plus a fifteen-field completeness gate enforced by database triggers; (2)three-dimension emergence scoring with six-stage discovery chain and evidence counting; (3)multi-source pseudo-temporal skill differencing that labels added/removed/modified requirements via reproducible $\Delta_s$; and (4)five-dimension matching with weights $0.42/0.24/0.14/0.10/0.10$,a $0.65$ transfer cap;(5)an $|S|\!\ge\!2$ evidence gate. Online evaluation shows JD parsing, resume extraction, and matching accuracies all at or above 90%; emerging-role F1 is 0.824 and evolution–expert Kappa is 0.76. Formulas, codeexcerpts, baselines, and ablations are reported so that claims remain checkable.

Keywords

The Job–Competency Knowledge Graph; Dynamic Evolution Analysis; Multi-source Data Governance; Emerging Job Discovery; Human-Job Matching; Retrieval-Augmented Generation; Hallucination Control

References

[1] Kleppmann M. Designing data-intensive applications: The big ideas behind reliable, scalable, and maintainable systems. Sebastopol: O'Reilly Media, 2017. [2] Batini C, Scannapieco M. Data and information quality: Dimensions, principles and techniques. Cham: Springer, 2016. [3] Peng C, Xia F, Naseriparsa M, et al. Knowledge Graphs: Opportunities and Challenges. Artificial Intelligence Review, 2023, 56: 13071-13102. DOI: 10.1007/s10462-023-10465-9. [4] Goyal N, Kalra J, Sharma C, et al. JobXMLC: EXtreme Multi-Label Classification of Job Skills with Graph Neural Networks//Findings of the Association for Computational Linguistics: EACL 2023. Dubrovnik, Croatia: Association for Computational Linguistics, 2023: 2181-2191. DOI: 10. 18653/v1/2023. Findings - eacl. 163. [5] Seif A, Toh S, Lee H K. A Dynamic Jobs-Skills Knowledge Graph//RecSys in HR'24: The 4th Workshop on Recommender Systems for Human Resources, in conjunction with the 18th ACM Conference on Recommender Systems. Bari, Italy, 2024. [6] Zhang Q, Gui T, Zheng R, et al. A review of research on named entity recognition. Journal of Software, 2018, 29(3): 812-831. [7] Lewis P, Perez E, Piktus A, et al. Retrieval-augmented generation for knowledge-intensive NLP tasks//Advances in Neural Information Processing Systems 33 (NeurIPS 2020). 2020: 9459-9474. [8] Li Y, Gao J, Meng C, et al. A survey on truth discovery. ACM SIGKDD Explorations Newsletter, 2016, 17(2): 1-16. [9] Y F Feng, H Hu, S H Ying, Xingliang Hou, Shiquan Liu, Mingyuan Yang, Junchang Li, Shaoyi Du, Nanning Zheng, Han Hu, and Yue Gao. Hyper-RAG: combating LLM hallucinations using hypergraph-driven retrieval-augmented generation. Nature Communications, vol. 17, Article 5778, 2026. DOI: 10. 1038/s41467-026-71411-1. [10] Nathan H, Marco S, Lu C. Differentiable Conformal Training for LLM Reasoning Factuality. In Proceedings of the 43rd International Conference on Machine Learning (ICML 2026), Poster #2501, 2026. [11] Hu X Y, Huang J, Lin Z R, et al. AI ethics in education: conceptual framework, cognitive state, and risk prevention. Modern Distance Education Research, 2022, 34(2): 21-29.