Logo PTI Logo FedCSIS

Communication Papers of the 21st Conference on Computer Science and Intelligence Systems (FedCSIS)

Annals of Computer Science and Information Systems, Volume 49

Semantic Matching of IT Vocational Training Offers with the ESCO Ontology

,

DOI: http://dx.doi.org/10.15439/2026F8838

Citation: Katerina Kostadinovska,

Full text

Abstract. This paper presents a pipeline for semantically matching IT vocational training offers from the KURSNet database to skills defined in the ESCO ontology. The proposed approach combines semantic IT filtering, LLM-based keyword cleaning, and embedding-based matching using Sentence Trans former models. Applied to a dataset of 6,128 IT training courses with 16,372 unique keywords, the pipeline successfully maps 63.4\% of keywords to ESCO skills at a cosine similarity threshold of 0.75. Evaluation against a manually annotated gold standard of 1,000 keyword-ESCO pairs yields a precision of 61.8\%, with an inter-annotator agreement of κ = 0.91. The results demonstrate that embedding-based methods are suitable for bridging semi-structured national training data with standardized European competency frameworks.

References

  1. OECD, Getting Skills Right: Future-Ready Adult Learning Systems. OECD Publishing, 2019.
  2. Bundesagentur für Arbeit, “Eingabehilfe für die erfassung von weiterbildungsangeboten in kursnet,” Bundesagentur für Arbeit, Nürnberg, Tech. Rep., 2025, stand: 31.03.2025.
  3. European Commission, ESCO Handbook: European Skills, Competences, Qualifications and Occupations, 2nd ed. Luxembourg: Publications Office of the European Union, 2019.
  4. N. Reimers and I. Gurevych, “Sentence-bert: Sentence embeddings using siamese bert-networks,” in EMNLP, 2019.
  5. N. Reimers and I. Gurevych, “Making monolingual sentence embeddings multilingual using knowledge distillation,” in Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Linguistics, 2020. [Online]. Available: https://arxiv.org/abs/2004.09813
  6. C. Bizer, T. Heath, and T. Berners-Lee, “Linked data – the story so far,” International Journal on Semantic Web and Information Systems, vol. 5, no. 3, pp. 1–22, 2009.
  7. World Wide Web Consortium (W3C), “Skos simple knowledge organization system reference,” https://www.w3.org/TR/skos-reference/, 2009, w3C Recommendation.
  8. T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient estimation of word representations in vector space,” in ICLR, 2013.
  9. C. D. Manning, P. Raghavan, and H. Schütze, Introduction to Information Retrieval. Cambridge: Cambridge University Press, 2008.
  10. T. Petrican and O. Stan, “Ontology-based skill matching algorithms,” in Proceedings of CSCS, 2015.
  11. S. Paudel and S. Shakya, “Ontology based job-candidate matching using skill sets,” in IOE Graduate Conference, 2017.
  12. L. Fernández-Sanz et al., “e-skills match: A framework for mapping skills into esco,” Computer Standards & Interfaces, vol. 51, pp. 30–42, 2017.
  13. L. Snijder et al., “Advancing ontology alignment in the labor market,” in AAAI-MAKE 2024, 2024.
  14. J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “Bert: Pre-training of deep bidirectional transformers for language understanding,” in NAACLHLT, 2019.
  15. N. Bocharova et al., “Vacancybert: A domain-adapted language model for job matching,” Applied Sciences, 2021.
  16. A. Bhola et al., “Extracting skills from job descriptions using embeddings,” in Proceedings of the 12th Language Resources and Evaluation Conference (LREC 2020), 2020.
  17. T. B. Brown et al., “Language models are few-shot learners,” in Advances in Neural Information Processing Systems (NeurIPS), 2020. [Online]. Available: https://arxiv.org/abs/2005.14165
  18. P. Liu, W. Yuan, J. Fu, Z. Jiang, H. Hayashi, and G. Neubig, “Pretrain, prompt, and predict: A systematic survey of prompting methods in natural language processing,” ACM Computing Surveys, vol. 55, no. 9, pp. 1–35, 2023.
  19. J. Cohen, “A coefficient of agreement for nominal scales,” Educational and Psychological Measurement, vol. 20, no. 1, pp. 37–46, 1960.
  20. J. R. Landis and G. G. Koch, “The measurement of observer agreement for categorical data,” Biometrics, vol. 33, no. 1, pp. 159–174, 1977.
  21. OECD, OECD Skills Outlook 2021: Learning for Life. Paris: OECD Publishing, 2021.
  22. N. Thakur, N. Reimers, A. Rücklé, A. Srivastava, and I. Gurevych, “BEIR: A heterogeneous benchmark for zero-shot evaluation of information retrieval models,” in Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks, 2021. [Online]. Available: https://arxiv.org/abs/2104.08663