딥러닝기반 텍스트 분석을 통한 직업분류시스템 구축에 관한 연구
A study of the establishment of a job classification system with deep learning-based online job posting text analysis
2
조회
0
다운로드
3
인용
0
참고문헌
키워드
초록
본 연구의 목적은 온라인구인공고 텍스트 데이터를 활용하여 해당 일자리의 직종을 판별하는 분류 모델을 생성해 평가하는 것이다. 워크넷 온라인구인공고(OJPs) 텍스트 자료에 딥러닝 기계학습 기법을 적용하여 자동으로 직업을 판별하는 것이다. 텍스트 자동 분류를 위한 기계학습 기법이 규칙기반 모델에서 인공신경망 모델로 전환하는 연구 흐름을 반영하고, 대규모의 온라인구인공고 자료와 텍스트의 문맥적 의미를 잘 다룰 수 있다는 점을 고려하여, 인공신경망의 최신 모델인 Bi-LSTM과 KoBERT 모델을 적용하였다. 1999-2021년 간의 워크넷 구인공고 데이터 800만 개에 모델을 적용한 결과, 0.62-0.82 정도의 매칭 정확도를 달성했다. 특히, 직무 기술(job description)이 특수하고 정확한 전문직에서 높은 정확도를 달성했다.
Abstract
The purpose of this study is to create a classification model that can identify the type of job by using online job posting text data and evaluate the performance of the model. By applying the latest deep learning machine learning method to Work-Net online job postings(OJPs) text data, it is to automatically determine the occupat ional code of the OJPs. Considering the research trends shifting from a rule-based model to an artificial neural network model, and the merit of handling large-scale online job posting materials and the contextual meaning of text well, the latest models of artificial neural networks, Bi-LSTM and KoBERT models, were applied. As a result of applying the model to 8 million text data of employment insurance Work-Net job posting data from 1999 to 2001, matching accuracy of 0.62 to 0.82 was achieved. The result is not very high performance, but it is generally judged to be a model that can determine the occupation. In particular, high accuracy was achieved in professions where job descriptions were specific and precise. Although it is not yet perfect for practical use, it is expected that the performance of the automatic occupational classification system will improve in the future when recruitment practices into the job-type labor market change and more precise data pre-processing and model applications are made.
참고문헌 (21)
- 오현주, "취업취약계층의 취업장애요인 유형 분류에 관한 연구: 취업성공패키지 참여자를 중심으로", 한국진로창업경영학회지, 4(1), pp.71~100 (2020), 한국진로·창업경영학회[DOI]
- 오성욱, "대학생의 취업준비행동이 졸업 후 첫 직업에 대한 만족요인들 간의 인과관계", 한국진로창업경영학회지, 4(1), pp.23~50 (2020), 한국진로·창업경영학회[DOI]
- 우찬균, "딥러닝 기반 한국 표준 산업분류 자동분류 모델 비교", 한국정보처리학회 학술대회논문집, 27(1), pp.516~518 (2020)
- 임정우, "딥러닝 기법을 활용한 산업", 직업 자동코딩 시스템, 2021, pp.12(4), 23~30[DOI]
- Acemoglu, D., "AI and jobs: Evidence from online vacancies (No. w28257)", National Bureau of Economic Research, ~ (2020)
- Boselli, R., "WoLMIS : a labor market intelligence system for classifying web job vacancies", Journal of Intelligent Information Systems, 51(3), pp.477~502 (2018)
- Boselli, R., "Classifying online job advertisements through machine learning", Future Generation Computer Systems, 86, pp.319~328 (2018)
- Boselli, R., "Joint European Conference on Machine Learning and Knowledge Discovery in Database" (2017), Springer
- No longer qualified? Changes in the supply and demand for skills within occupations, "Burke, M. A.", Federal Reserve Bank of Boston, ~ (2020)
- Calanca, F., "Responsible team players wanted : an analysis of soft skill requirements in job advertisements", EPJ Data Science, 8(1), pp.1~20 (2019)
- Choi, I. H., "A Study of the Classification of IT Jobs Using LSTM and LIME", The 9th International Conference on Smart Media and Applications (2020), Association for Computing Machinery
- Colombo, E., "AI meets labor market : Exploring the link between automation and skills", Information Economics and Policy, 47, pp.27~37 (2019)
- Das, S., "Forecasting Task-Shares and Characterizing Occupational Change across Industry Sectors", Harvard CRCS Workshop on AI for Social Good (2020), CRCS
- Deming, D. J., "Earnings dynamics, changing job skills, and STEM careers", The Quarterly Journal of Economics, 135(4), pp.1965~2005 (2020)
- Gnehm, A. S., "Text zoning and classification for job advertisements in German, French and English", Proceedings of the Fourth Workshop on Natural Language Processing and Computational Social Science (2020), Association for Computational Linguistics
- Hershbein, B., "Do recessions accelerate routine-biased technological change? Evidence from vacancy postings", American Economic Review, 108(7), pp.1737~1772 (2018)
- Hoang, P., "Large-scale occupational skills normalization for online recruitment", AI Magazine, 39(1), pp.5~14 (2018)
- Marrara, S., "A language modelling approach for discovering novel labour market occupations from the web", Proceedings of the International Conference on Web Intelligence (2017), Association for Computing Machiner
- Tamburri, D. A., "DataOps for societal intelligence: A data pipeline for labor market skills extraction and matching", 2020 IEEE 21st International Conference on Information Reuse and Integration for Data Science(IRI) (2020), IEEE
- Transforming naturally occurring text data into economic statistics: The case of online job vacancy postings, "Turrell, A.", National Bureau of Economic Research, ~, pp.WP No. w25837 (2019)
- Xu, H., "JCTC : A Large Job posting Corpus for Text Classification", arXiv preprint arXiv : 1705, ~ (2017)