소셜 데이터 마이닝을 통한 인터 카테고리 브랜드 맵의 생성

다국어 초록 (Multilingual Abstract)

As social media in diverse forms are evolving, it is quite common to witness the big consequence of a single user’s on-line opinion both positive and negative. Social media, a tool for expressing and exchanging ideas and thoughts between people in general, is being regarded as an essential medium for collecting and monitoring public opinion. Normal people post various aspects of their everyday lives including preferences for
products and services in market.
This thesis attempts to extract users’ interests and preferences toward multiple brands in order to identify and embody the groupings of brands across conventional product categories. This phenomenon is called inter-category brand map. Being based on the concept of constructed market suggested by economic sociologists, this study presents a framework for deriving inter-category brand map by applying natural language processing and text mining technologies on large-scale social media data. The crucial factor is the similarities or distances between product categories and brands
measured by mention patterns extracted from a huge pool of blog posts.
This thesis argues and proves brands that are repeatedly mentioned by multiple users in a given period of time tend to have greater influences in market and some of the brands are frequently mentioned together even though they belong to different product categories forming a mentally constructed dynamic category.
The results of the experiments that deal with user preferences and inter-category and inter-brand affinities unveil the possibility of using inter-category brand map in various real-world marketing activities including alliance and/or collaboration marketing. It would be also possible to use inter-category map in developing product bundling and building organizational portfolio.
If the framework and methods proposed in this study is further improved in future work, it will provide even more powerful tool for monitoring and predicting the dynamic trend for market shift.

목차 (Table of Contents)

• 제 1 장 서론 1
• 1.1 연구의 배경 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
• 1.2 연구의 목적과 기본 가설 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
• 1.3 논문의 구성 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
• 제 2 장 기존 연구 및 이론적 토대 10
• 2.1 기존 연구 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
• 2.1.1 온라인 구전 및 확산 관련 연구 . . . . . . . . . . . . . . . . . . . . . . . . . 10
• 2.1.2 온라인 네트워크 상에서의 지식 공유 관련 연구 . . . . . . . . . . . . . . . . 12
• 2.1.3 사회 연결망 분석 기법을 이용한 온라인 네트워크 분석 . . . . . . . . . . . 13
• 2.1.4 기존 연구의 한계 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
• 2.2 이론적 토대 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
• 2.2.1 복잡한 사용자들의 선호 형성 . . . . . . . . . . . . . . . . . . . . . . . . . . 15
• 2.2.2 시장 간의 인지적 경계 형성 . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
• 2.2.3 시장 간 연결망과 소비자의 범주화 . . . . . . . . . . . . . . . . . . . . . . . 18
• 2.2.4 경제사회학과 범주 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
• 2.2.5 사용자들의 선호와 다중 범주 . . . . . . . . . . . . . . . . . . . . . . . . . . 21
• 2.2.6 주관성 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
• 2.2.7 통합 모형의 제기 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
• 2.3 연구의 시사점 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
• 제 3 장 자료의 구축과 검증 26
• 3.1 소셜 데이터의 수집 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
• 3.2 소셜 데이터의 가공 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
• 3.3 언급 자료의 구축 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
• 3.3.1 언급 행렬 생성 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
• 3.3.2 언급 유사도 측정 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
• 3.4 언급 자료의 검증 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
• 3.4.1 분석 대상 상품군 설정 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
• 3.4.2 초기 변수의 설정 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
• 3.4.3 통계 분석 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
• 3.5 요약 및 시사점 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
• 제 4 장 실험 및 분석 52
• 4.1 범주 간 유사도와 브랜드 선호의 연관성 분석 . . . . . . . . . . . . . . . . . . . . 52
• 4.1.1 브랜드 및 범주 간 유사도 분석 . . . . . . . . . . . . . . . . . . . . . . . . . 52
• 4.1.2 브랜드 선호 분석 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
• 4.2 브랜드 및 범주 간 친화도의 분석 . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
• 4.2.1 분석 자료 구축 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
• 4.2.2 브랜드 및 범주 간 거리의 측정 . . . . . . . . . . . . . . . . . . . . . . . . . 60
• 4.2.3 브랜드 간 친화도 분석 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
• 4.2.4 브랜드별 친화도 분석 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
• 4.2.5 범주 간 친화도 분석 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
• 4.3 요약 및 시사점 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
• 제 5 장 결론 80
• 참고 문헌 83
• Abstract 93