@article{ART003302942},
author={ChanJun Park and Seungil Choi},
title={A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions},
journal={Industry Promotion Research},
issn={2466-1139},
year={2026},
volume={11},
number={1},
pages={109-119},
doi={10.21186/IPR.2026.11.1.109}
TY - JOUR
AU - ChanJun Park
AU - Seungil Choi
TI - A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions
JO - Industry Promotion Research
PY - 2026
VL - 11
IS - 1
PB - Industrial Promotion Institute
SP - 109
EP - 119
SN - 2466-1139
AB - Each year, eligibility screening is conducted for more than 840,000 applicants for the National Technical Qualification examinations. This screening helps ensure that the qualification tests effectively verify the skills and knowledge required for actual job performance, and it is an important procedure for guaranteeing that certified individuals possess competencies appropriate to the qualification. However, the screening period is very short—within five days—and the process is carried out largely through manual work by reviewers. As a result, there is a risk that the consistency and quality of screening decisions may vary depending on each reviewer’s level of proficiency and work experience. To address this issue, this study aimed to improve the consistency and quality of eligibility screening by automating the process using state-of-the-art AI technologies, including large language models (LLMs) and retrieval-augmented generation (RAG). RAG has recently been recognized as an effective approach for mitigating hallucination problems in LLMs and for improving accuracy. In the system performance test, the accuracy showed top-1 0.72 and top-3 0.83, and the weighted average performance test top-1 showed precision 0.73, recall 0.72, and F1-score 0.72, confirming excellent performance. In addition, it is significant to present a methodology for judgment automation based on the accumulated screening data, laying the foundation for the innovative development of the screening process. In addition, solutions were proposed for the limitations of development, such as the need for information security in the public sector.
KW - Automated eligibility screening;Large Language Model;LLM;Retrieval Augmented Generation;RAG
DO - 10.21186/IPR.2026.11.1.109
ER -
ChanJun Park and Seungil Choi. (2026). A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions. Industry Promotion Research, 11(1), 109-119.
ChanJun Park and Seungil Choi. 2026, "A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions", Industry Promotion Research, vol.11, no.1 pp.109-119. Available from: doi:10.21186/IPR.2026.11.1.109
ChanJun Park, Seungil Choi "A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions" Industry Promotion Research 11.1 pp.109-119 (2026) : 109.
ChanJun Park, Seungil Choi. A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions. 2026; 11(1), 109-119. Available from: doi:10.21186/IPR.2026.11.1.109
ChanJun Park and Seungil Choi. "A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions" Industry Promotion Research 11, no.1 (2026) : 109-119.doi: 10.21186/IPR.2026.11.1.109
ChanJun Park; Seungil Choi. A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions. Industry Promotion Research, 11(1), 109-119. doi: 10.21186/IPR.2026.11.1.109
ChanJun Park; Seungil Choi. A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions. Industry Promotion Research. 2026; 11(1) 109-119. doi: 10.21186/IPR.2026.11.1.109
ChanJun Park, Seungil Choi. A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions. 2026; 11(1), 109-119. Available from: doi:10.21186/IPR.2026.11.1.109
ChanJun Park and Seungil Choi. "A study on the Automating Eligibility Screening for National Technical Qualification Examinations Based on Retrieval-Augmented Generation (RAG): Focusing on Automated Classification of Job Descriptions" Industry Promotion Research 11, no.1 (2026) : 109-119.doi: 10.21186/IPR.2026.11.1.109
