隨著社會網絡和電子商務的快速發展，潛在消費者們在做出購買決定之前，經常從各種線上資源中尋求其他顧客所撰寫與產品相關資訊的評論及看法。透過web 2.0的發展，客戶能夠針對各種產品提供自己意見或討論板中與其他顧客溝通，交流和分享彼此的使用經驗。有鑑於豐富的線上資源，本研究欲提出一關聯分析方法結合評論內容和社交網絡角色分析以串聯社群網絡中社會角色與評論行為的關係。本研究將藉由生成social-preference矩陣，包含每個使用者的產品偏好以及他/她的社會角色，，透過關聯規則技術來串接兩項線上資源間的關係。分析結果表明，特定的社會角色有不同的產品偏好以及各角色在市場中的影響力。我們所提出的方法可以為公司提供有效的見解，作為推動行銷和產品客製化之基礎，協助企業了解顧客們在社群網路中所扮演的社交角色以及對於自家服務或產品的偏好差異。

With the rapid development of social Web and electronic commerce, the potential consumers intend to seek for contents with product-related information before making a purchase decision from various online resources. Among these resources, customer reviews provide the information of the opinion for various product features. In addition, the discussion boards consist of peer communication to exchange and share their experiences with each other. Therefore, in this study, we aim to propose a method that combines content and social network analysis to discover the valuable patterns for market analysis. The proposed method generates the social-preference matrix which includes each user’s product preference and his/her social role defined by clustering to bridge the relationships by employing association rules techniques. The results show that different product features are concerned with the specific social role. Our method can provide relevant insights for companies to launch effective marketing and design product customization and for consumers easily to get ideas about their purchase decisions. This study particularly investigates customer review and social network and explores how two social resources associated with each other to improve the insight into each user’s preference with a social role they act.

TABLE OF CONTENTS
CHAPTER1 – Introduction 1
1.1 Background and Motivation 1
1.2 Research Purpose 3
1.3 Contribution 4
1.4 Thesis Organization 4
CHAPTER2 – Literature Review 5
2.1 Managerial Aspect 5
2.2 Technical Aspect 8
CHAPTER3 – Problem Definition 14
3.1 Preliminaries 14
3.2 Problem Description 14
CHAPTER4 – Methodology 16
4.1 Overall Process Architecture 16
4.2 Data collection stage 17
4.3 Social activity analysis 18
4.4 User-product review analysis 23
4.5 Association Rules 27
CHAPTER5 – Case study 30
5.1 Dataset Description 30
5.2 Data Preprocessing 31
5.3 Association Rules Results 31
CHAPTER6 – Conclusions 49
References 52

LIST OF FIGURES
Figure 3‑1. Problem Definition 14
Figure 4‑1. Overall Process of Our Research 16
Figure 4‑2. Rank features by importance 20
Figure 4‑3. The RMSE of subsets of features 21
Figure 4‑4. The optimal cluster determined by Gap Statistic 21
Figure 4‑5. The cluster center of each social role after feature selection 22
Figure 4‑6. The determination of number of topics by Coherence score 25
Figure 4‑7. Average topic proportion within document by year 25
Figure 5‑1.Graph-based visualization in each cluster in 2012 33
Figure 5‑2. Association rules between top features in Grunts in 2012 35
Figure 5‑3. Association rules between top preferences in Taciturns in 2012 37
Figure 5‑4. Association rules between topic preferences in Supporters in 2012 37
Figure 5‑5. Association rules between topic preferences in Popular Initiators in 2012 38
Figure 5‑7. Association rules between top preferences in Taciturns in 2013 41
Figure 5‑8. Association rules between top preferences in Supporters in 2013 42
Figure 5‑9. Association rules between top preferences in Popular Initiators in 2013 42
Figure 5‑11. Association rules between top preferences in Grunts in 2014 45
Figure 5‑12. Association rules between top preferences in Taciturns in 2014 45
Figure 5‑13. Association rules between top preferences in Supporters in 2014 46
Figure 5‑14. Association rules between top preferences in Popular Initiators in 2014 46

LIST OF TABLES
Table 4‑1. The definitions of each feature in user's vector 19
Table 4‑2. Rules for differentiating user clusters 22
Table 4‑3. The Numbers of Roles in Each Cluster 23
Table 4‑4. Top keywords in each topic 26
Table 4‑5. Product review and its information extraction 27
Table 5‑1. Association rules between topic features and social role in 2012 33
Table 5‑2. Association rules between topic features and social role in 2013 39
Table 5‑3. Association rules between topic features and social role in 2014 43

Chan, J., & Hayes, C. (2010). Decomposing Discussion Forums using User Roles. In In Proceedings of the International Web Science Conference (Vol. 112, pp. 215–218).
Chaney, A. J. B., Blei, D. M., & Eliassi-rad, T. (2015). A Probabilistic Model for Using Social Networks in Personalized Item Recommendation. In Proceedings of the 9th ACM Conference on Recommender Systems (pp. 43–50).
Cheng, L. C., Ke, Z. H., & Shiue, B. M. (2011). Detecting changes of opinion from customer reviews. In Fuzzy Systems and Knowledge Discovery (FSKD), 2011 Eighth International Conference (Vol. 3, pp. 1798–1802).
Duan, J., Zeng, J., & Luo, B. (2014). Identification of opinion leaders based on user clustering and sentiment analysis. In Proceedings of the 2014 IEEE/WIC/ACM International Joint Conferences on Web Intelligence (WI) and Intelligent Agent Technologies (IAT) (Vol. 1, pp. 377–383).
Emoto, M. (2016). Method for Extraction of Purchase Behavior and Product Character Using Dynamic Topic Model. Data Mining Workshops (ICDMW), 2016 IEEE 16th International Conference, 778–782.
Guo, S., Wang, M., & Leskovec, J. (2011). The Role of Social Networks in Online Shopping: Information Passing, Price of Trust, and Consumer Choice. ACM Conference on Electronic Commerce, 157–166.
He, W., Zha, S., & Li, L. (2013). Social media competitive analysis and text mining: A case study in the pizza industry. International Journal of Information Management, 33, 464–472.
Kim, Y. A., & Srivastava, J. (2007). Impact of Social Influence in E-Commerce Decision Making. In Proceedings of the ninth international conference on Electronic commerce (pp. 293–302).
Lackermair, G., Kailer, D., & Kanmaz, K. (2013). Importance of Online Product Reviews from a Consumer ’ s Perspective. Advances in Economics and Business, 1(1), 1–5.
Li, S., Nahar, K., & Fung, B. C. M. (2013). Product customization of tablet computers based on the information of online reviews by customers. Journal of Intelligent Manufacturing, 26(1), 97–110.
Maharani, W., & Gozali, A. A. (2015). Collaborative Social Network Analysis and Content-based Approach to Improve the Marketing Strategy of SMEs in Indonesia. Procedia Computer Science, 59, 373–381.
McAuley, J., Pandey, R., & Leskovec, J. (2015). Inferring Networks of Substitutable and Complementary Products. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 785–794).
McAuley, J., Targett, C., Shi, Q., & Van Den Hengel, A. (2015). Image-based Recommendations on Styles and Substitutes. In Proceedings of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval (pp. 43–52).
McAuley, J., & Yang, A. (2016). Addressing Complex and Subjective Product-Related Queries with Customer Reviews. In Proceedings of the 25th International Conference on World Wide Web (pp. 625–635).
Morzy, M. (2009). On mining and social role discovery in internet forums. Social Informatics, 2009. SOCINFO’09. International Workshop, 74–79.
Moser, C., Groenewegen, P., & Huysman, M. (2013). Extending social network analysis with discourse analysis: Combining relational with interpretive data. The Influence of Technology on Social Network Analysis and Mining, 6(May), 547–561.
Mostafa, M. M. (2013). More than words: Social networks’ text mining for consumer brand sentiments. Expert Systems with Applications, 40(10), 4241–4251.
Najafabadi, M. K., Mahrin, M. N., Chuprat, S., & Sarkan, H. M. (2017). Improving the accuracy of collaborative filtering recommendations using clustering and association rules mining on implicit data. Computers in Human Behavior, 67, 113–128.
Ng, C. S. P. (2013). Intention to purchase on social commerce websites across cultures: A cross-regional study. Information & Management, 50(8), 609–620.
Thackeray, R., Neiger, B. L., Hanson, C. L., & McKenzie, J. F. (2008). Enhancing promotional strategies within social marketing programs: use of Web 2.0 social media. Health promotion practice. Health Promotion Practice, 9(4), 338–343.
Tibshirani, R., Walther, G., & Hastie, T. (2001). Estimating the number of clusters in a data set via the gap statistic. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 63(2), 411–423.
Trusov, M., Bucklin, R. E., & Pauwels, K. (2009). Effects of Word-of-Mouth Versus Traditional Marketing : Findings from an Internet Social Networking Site. Journal of Marketing, 73(5), 90–102.
Wang, Z., Sun, L., Zhu, W., Yang, S., Li, H., & Wu, D. (2013). Joint social and content recommendation for user-generated videos in online social network. IEEE Transactions on Multimedia, 15(3), 698–709.
Xianghua, F., Guo, L., Yanyan, G., & Zhiqiang, W. (2013). Multi-aspect sentiment analysis for Chinese online social reviews based on topic modeling and HowNet lexicon. Knowledge-Based Systems, 37, 186–195.
Zhang, J., & Piramuthu, S. (2016). Product recommendation with latent review topics. Information Systems Frontiers, 1–9.