텍스트 마이닝과 토픽 모델링을 이용한 성조숙증 관련 인터넷 건강상담 분석

• Published in: 보건교육건강증진학회지 Vol. 37; no. 3; pp. 71 - 84
• Main Authors: 박수잔(Susan Park), 서영균(Young-Kyoon Suh), 김재성(Jaesung Kim), 권진원(Jin-Won Kwon)
• Format: Journal Article
• Language: Korean
• Published: 한국보건교육건강증진학회 01.09.2020
• Subjects: 예방의학; health education; topic modeling; text mining; precocious puberty
• ISSN: 1229-4128; 2635-5302
• DOI: 10.14367/kjhep.2020.37.3.71

Objectives: This study aimed to investigate the characteristics of the demand for precocious puberty-related health information using the text data of online health counseling through the techniques of text mining and topic modeling. Methods: The input data of this study were question and answer documents that were searched for Naver Jisik-iN using the keywords 'precocious puberty'. The documents were automatically collected using a web crawler from January 2017 to June 2019. We performed text pre-processing, after which all nouns were extracted using a morphological analyzer. LDA(Latent Dirichlet Allocation) topic modeling was performed to find hidden themes in the text. Results: A majority of questions were submitted during the school vacation. Approximately half of all questions were answered by doctors. We used LDA topic modeling to yield three themes regarding precocious puberty-related questions: 1) asking for expert opinion on diagnosis or hospital selection, 2) questions regarding diagnostic tests and treatment, and 3) questions considering obesity and/or diet, as well as timing of puberty onset. Conclusion: This study identified the major precocious puberty-related concerns. In light of the increasing incidence of precocious puberty, individuals with this condition, as well as parents, should be provided with appropriate education to promote awareness and management. KCI Citation Count: 0