머신러닝 언어모델을 활용한 문법적 연어 추출 방법론 연구 - 명사를 중심어로 하는 문법적 연어를 중심으로 -
A study on extracting grammatical collocations using machine learning language models - Focusing on grammatical collocations whose nodes are nouns -
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This study aimed to select features for identifying Korean grammatical collocations and explored the feasibility of classifying grammatical collocations using a machine-learning model. Candidate grammatical collocations whose nodes are nouns were extracted from the Sejong Corpus. Six features were identified: the frequency of lexical chains, the entropy of adjacent words in lexical chains, the average distance and variance between components of lexical chains, and the entropy of preceding and following elements of lexical chains whose nodes are nouns. To address the class imbalance in the training data, eight sampling techniques were applied. Both the original and sampled datasets were trained using XGBoost to develop a classification model and evaluate its performance. The results indicated that the model trained on data sampled using a combination of SMOTE and ENN exhibited the highest accuracy and better classified the minority class of grammatical collocations. The most influential features were 'frequency', 'minimum entropy', and 'variance of average distance'. However, limitations were identified, as the frequency and information of adjacent words alone were insufficient to fully capture the contextual meaning and grammatical characteristics of collocations. Future research should utilize transformer-based pre-trained models and embedding techniques to extract features that reflect the contextual meaning and functions of grammatical collocations. This approach is expected to facilitate an objective review of grammatical collocation lists for Korean language education and establish unified criteria for identifying grammatical collocations.
Abstract
This study aimed to select features for identifying Korean grammatical collocations and explored the feasibility of classifying grammatical collocations using a machine-learning model. Candidate grammatical collocations whose nodes are nouns were extracted from the Sejong Corpus. Six features were identified: the frequency of lexical chains, the entropy of adjacent words in lexical chains, the average distance and variance between components of lexical chains, and the entropy of preceding and following elements of lexical chains whose nodes are nouns. To address the class imbalance in the training data, eight sampling techniques were applied. Both the original and sampled datasets were trained using XGBoost to develop a classification model and evaluate its performance. The results indicated that the model trained on data sampled using a combination of SMOTE and ENN exhibited the highest accuracy and better classified the minority class of grammatical collocations. The most influential features were 'frequency', 'minimum entropy', and 'variance of average distance'. However, limitations were identified, as the frequency and information of adjacent words alone were insufficient to fully capture the contextual meaning and grammatical characteristics of collocations. Future research should utilize transformer-based pre-trained models and embedding techniques to extract features that reflect the contextual meaning and functions of grammatical collocations. This approach is expected to facilitate an objective review of grammatical collocation lists for Korean language education and establish unified criteria for identifying grammatical collocations.
참고문헌 (46)
- 강현화, "한국어교육 문법: 자료편" (2016), 한글파크
- 경희한국어 교재편찬위원회, "《경희한국어》초급1~고급2" (2002), 경희대학교 출판문화원
- 고려대학교 한국어센터, "《고려대한국어》1A~4B" (2020), 고려대학교 출판문화원
- 국립국어원, "외국인을 위한 한국어 문법2" (2009), 커뮤니케이션북스
- 국제 통용 한국어 표준 교육과정 적용 연구, "김중섭", 국립국어원, ~ (2017)
- 남길임, "학술 텍스트에 나타난 핵심 구문의 추출-국어국문학 학술논문 말뭉치를 중심으로-", 어문론총, 60, pp.65~92 (2014), 한국문학언어학회
- 박경미, "엔트로피를 이용한 한국어 연어 추출", 석사, pp.2002
- 박병선, "한국어 계량적 연구 방법론" (2005), 역락
- 박형진, "‘문법적 연어’에 대한 고찰", 열린정신 인문학연구, 10(1), pp.41~57 (2009), 원광대학교 인문학연구소
- 백봉자, "외국어로서의 한국어 문법 사전" (2009), 도서출판 하우
- 서강대학교 한국어교육원, "《NEW 서강 한국어》1A~6" (2008), 서강대학교 한국어교육원
- 서울대학교 언어교육원, "《서울대 한국어》1A~6B" (2013), 문진미디어
- 신효필, "언어 자료의 통계 분석과 관련된 몇 가지 고려사항들", 어학연구, 41, pp.655~682 (2005), 서울대학교 어학연구소
- 신효필, "연어의 통계적 접근을 통한 로그우도비 중심의 연어 검증", 언어학, 47, pp.107~138 (2007), 한국언어학회
- 연세대학교 한국어학당, "《연세 한국어》1-1~6-2" (2013), 연세대학교 출판부
- 이 진, "한국어 문법적 연어 추출 방법론 연구", 박사, pp.2022
- 이두행, "통계에 기반한 한국어 연어 결합 측정의 평가", 석사, pp.2011
- 이은하, "한국어 구어 말뭉치를 대상으로 한 연어구성 추출 방법들의 비교 : 재현가능 연구", 언어와 언어학, 70, pp.497~539 (2016), 한국외국어대학교 언어연구소
- 이화여자대학교 언어교육원, "《이화 한국어》1-1~6" (2010), 이화여자대학교 출판부
- 이희자, "한국어 교육용 조사·어미 사전" (2001), 한국문화사
- 임근석, "현대 국어의 어휘적 연어 연구", 석사, pp.2002
- 임근석, "한국어 연어 연구", 박사, pp.2006
- 임근석, "통계적 방법을 이용한 문법적 연어 후보 추출", 한국어학, 45, pp.305~333 (2009), 한국어학회
- 임근석, "한국어 연어 연구의 전개와 쟁점에 대하여", 국어학, 61, pp.359~387 (2011), 국어학회
- 임춘매, "공기어 분석을 통한 한국어 교육용 연어 목록 선정 : '체언+용언'형 구성을 중심으로", 박사, pp.2020
- 장석배, "한국어 정형표현 연구-대규모 말뭉치를 중심으로", 박사, pp.2015
- 정성훈, ""부사-용언"관계를 이용한 타당한 연어 검정 방법 검토 : 세종 코퍼스 구어자료를 중심으로", 한국어학, 66, pp.245~278 (2015), 한국어학회
- 주은석, "병렬말뭉치의 구축과 활용; 대규모 연어추출을 위한 통계적 기법 비교", 언어사실과 관점, 25, pp.189~210 (2010), 연세대학교 언어정보연구원
- 최준, "한국어의 정형화된 표현 연구", 담화와 인지, 17(2), pp.163~190 (2010), 담화·인지언어학회
- 한국외국어대학교 한국어문화교육원, "《외국인을 위한 한국어》1-1~6" (2016), 하우
- 허희정, "학술 텍스트 정형 표현 연구: 연결 표현과 종결 표현의 공기관계를 중심으로", 석사, pp.2016
- 홍종선, "한국어 연어 정보의 분석· 응용에 관한 연구", 한국어학, 11, pp.73~158 (2000), 한국어학회
- Abou Omar, K. B., "XGBoost and LGBM for Porto Seguro’s Kaggle challenge: A comparison" (2018), Preprint Semester Project
- Chen, T., "Xgboost: Reliable large-scale tree boosting system", Proceedings of the 22nd SIGKDD Conference on Knowledge Discovery and Data Mining, pp.13~17 (2015)
- Chen, T., "Xgboost: A scalable tree boosting system", Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, pp.785~794 (2016)
- Garner, J., "Beginning and intermediate L2 writer’s use of N-grams : an association measures study", International Review of Applied Linguistics in Language Teaching, 58(1), pp.51~74 (2020)
- Kilgarriff, A., "Collocationality (and how to measure it)", Proceedings of the Euralex Conference, pp.997~1004 (2006)
- Meng, Y., "What makes an online review more helpful : an interpretation framework using XGBoost and SHAP values", Journal of Theoretical and Applied Electronic Commerce Research, 16(3), pp.466~490 (2020)
- N. V. Chawla, "SMOTE : Synthetic Minority Over-sampling Technique", Journal of Artificial Intelligence Research, 16, pp.321~357 (2002)
- Pearce, D., "Synonymy in collocation extraction", Proceedings of the workshop on WordNet and other lexical resources, second meeting of the north american chapter of the association for computational linguistics, pp.41~46 (2001)
- Pearce, D., "A Comparative Evaluation of Collocation Extraction Techniques", LREC, pp.1530~1536 (2002)
- Pecina, P., "Combining association measures for collocation extraction", Proceedings of the COLING, 2006, 651~658
- Su, Q., "Association measures for collocation extraction : Automatic evaluation on a large-scale corpus", International Journal of Corpus Linguistics, 29(1), pp.59~86 (2024)
- Wermter, J., "Collocation extraction based on modifiability statistics", COLING 2004: Proceedings of the 20th International Conference on Computational Linguistics, pp.980~986 (2004)
- Wermter, J., "You can’t beat frequency(unless you use linguistic knowledge)–a qualitative evaluation of association measures for collocation and term extraction", Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, pp.785~792 (2006)
- Yang, S., "Machine learning for collocation identification", International Conference on Natural Language Processing and Knowledge Engineering Proceedings, pp.315~320 (2003)