This study applies the topic modeling technique, a prominent method in text mining, to the field of Japanese studies, aiming to address the challenges posed by the growing volume of textual data. To provide a deeper understanding of Korean interest in Japan, the research focuses on social media data generated by the ordinary users, rather than traditional mass media sources. Specifically, the study collected and constructed a corpus from ten years of Japan-related blog posts on the Korean platform Naver, which was then analyzed using topic modeling. The analysis revealed that these Japan-related blog posts could be broadly categorized into six topics. Based on the frequent keywords associated with each topic, these were identified as "Travel," "Economy," "Learning," "Shopping," "Popular Culture," and "Japanese Cuisine." The findings indicate that Koreans primarily view Japan as a travel destination, followed by its significance as an economic entity in international relations, a subject of study or destination for learning the Japanese language and culture, a shopping destination, a source of popular culture, and a provider of culinary experiences. These patterns demonstrate that topic modeling effectively identifies themes and consumption patterns, reflecting real-world social phenomena. This study underscores the potential of data mining, particularly topic modeling, in Japanese studies by enabling the analysis of large-scale textual data and deriving meaningful insights. However, limitations include the consolidation of data into a single set for analysis and the restriction of data collection to specific periods within each year. Future research should extend the data collection timeframe and incorporate data from other social media platforms to capture evolving trends in Korean interest in Japan over time.
1. 들어가며 2. 선행연구 및 문제제기 2.1 텍스트 마이닝(Text Mining) 2.2 토픽모델(Topic Model)과 소셜 빅데이터 분석 3. 조사 대상 및 연구방법 3.1 조사 대상 및 자료 수집 툴 3.2 데이터 전처리 3.3 데이터 분석 4. LDA 분석 및 결과 4.1 LDA에 의한 토픽 개수 설정 4.2 LDA에 의한 토픽 모델링 4.3. 토픽 분석 5. 나가며 【参考文献】 <要旨>
키워드
텍스트 마이닝토픽 모델LDA소셜 빅데이터블로그네이버Text MiningTopic ModelLatent Dirichlet AllocationSocial big dataBlogNaver
한국일본언어문화학회 [Japanese Language & Culture Association of Korea]
설립연도
2001
분야
인문학>일본어와문학
소개
본 학회는 일본어학 및 일본문학은 물론, 일본의 정치, 경제, 문화, 사회 등의 일본학 전반에 걸친 연구 및 일본의 언어, 문화를 매체로 한 한국과의 비교 연구를 대상으로 하고 있다. 본 학회는 회원들에게 연구 발표 및 정보 교환의 기회를 부여하고 나아가 한국에서의 바람직한 일본 연구 자세를 확립하는 것을 주된 목표로 하고 있다.