원문정보
Performance Evaluation of Open-source Korean NLP Tools : Focus on HR Data
초록
영어
This study evaluated the performance of open-source Korean natural language processing (NLP) tools with resume data, which include written texts to be analyzed for obtaining Human Resources (HR) insights. The results of this study showed that the overall performance of Korean morphological analyzers was satisfactory and Fasttext dealt with similar and rare vocabulary better than other algorithms of word embedding. Plus, TextRank and genism_summarizer successfully extracted the key phrases of a sentence. In contrast, Korean OCR packages produced poor results. Recently, an enormous amount of HR data such as resume and cover letter has been available, and if they are appropriately analyzed with NLP tools, the automation of HR work will be achieved.
목차
Ⅰ. 서론
Ⅱ. 자연언어처리(NLP)의 이해
2.1. 자연언어처리(NLP)란
2.2. 자연언어처리(NLP) 기술
2.3. 자연언어처리(NLP)의 활용
2.4. 자연언어처리(NLP)와 HR 연구
Ⅲ. 오픈소스 한국어 NLP 툴 성능분석
3.1. 형태소 분석기 성능 비교
3.2. 워드 임베딩(Word Embedding) 모델 성능 비교
3.3. 추천 시스템에서 요약 패키지 성능 비교
3.4. OCR (Optical Character Recognition) 패키지 성능 비교
Ⅳ. 결론
감사의 말
참고문헌