토픽 모델링을 활용한 수면 건강 소셜미디어 담론 분석 : 약물 기반과 비약물 기반 개입 비교
A Topic Modeling Analysis of Sleep Health Discourse on Social Media : A Comparison of Drug-based and Non-Drug-Based Interventions
Objectives: This study aims to analyze sleep health-related discourseon social media and to compare the characteristics of information-seeking behavior between drug-based and non-drug-based keyword groups using topic modeling techniques. Methods: A total of 4,434 blog posts were collected from Naver using five sleep-related keywords (“sleeping pills,” “Stilnox,” “melatonin,” “sleep routine,” and “sleep hygiene”) between March and October 2025. Latent Dirichlet Allocation (LDA) topic modeling was applied following morphological analysis. Chi-square tests were conducted to compare topic distributions between the drug-based (n=2,621) and non-drug-based (n=1,813) groups. Results: Nine topics were identified, with a coherence score of 0.5287. Significant differences in topic distributions between the two groups were observed (χ²=1815.85, p<0.001). The drug-based group was dominated by medication-related experience topics (53.6%), whereas the non-drug-based group focused primarily on general insomnia management (62.3%) and lifestyle habits (24.0%). Notably, only 1.3% of users in the non-drug-based group discussed medication-related topics. Conclusions: Social media users searching for drug-based and non-drug-based sleep interventions exhibit distinctly different information needs. These findings suggest the necessity for tailored health communication strategies targeting different groups of sleep health information seekers.
목차
Ⅰ. 서론 Ⅱ. 연구방법 1. 연구설계 2. 연구대상 3. 분석방법 Ⅲ. 연구결과 1. LDA 토픽 모델링 결과 2. 약물 기반 그룹과 비약물 기반 그룹의 토픽분포 비교 3. 키워드별 토픽 분포 특성 Ⅳ. 고찰 Ⅴ. 결론 REFERENCES