利用類別關聯規則探勘來協助顧客評論之情感分析

Mining Class Association Rules for Supporting Effective Sentiment Analysis in Consumer Reviews

施曉萍(Hsiao-Ping Shih)

指導教授 ： 楊錦生

元智大學/資訊學院/資訊管理學系/碩士(2012年)

https://doi.org/10.6838/YZU.2012.00215

摘要

Web 2.0應用的成功與普及，促成大量的使用者積極投入社群媒體 (Social Media) 的應用，例如，部落格、論壇、社群網路等。使用者在這些社群媒體內的行為與分享資訊，對企業而言是相當寶貴的資產，其可以協助企業更加了解廣大的使用者，以發展重要的商業智慧應用。在眾多的社群媒體與使用者產生的資料中，線上顧客評論是相當重要的一類。正確的顧客評論分析，可以提供顧客有價值的資訊，作出購買決策的重要決定。同時也可以幫助零售商或產品製造商了解顧客對其販售或生產的產品的看法與觀感，協助行銷策略的制定或產品的設計與改善。然而使用者發表產品評論的文章急劇成長，使得使用者 (例如，零售商、產品製造商、消費者) 需花費大量的時間和精力在顧客評論中去尋找有興趣的產品，並將之一一閱讀後再過濾出有幫助的資訊，這種透過手動來分析出顧客意見的正、負面情緒的方式，變得是困難又耗時的。因此本研究提出一個以規則為基礎的情感分析(Rule-based Sentiment Analysis)技術，其主要目的是自大量的顧客評論中，可自動萃取顧客表示的意見為何及針對的是產品的哪些屬性，不需要為每個產品種類的各個產品特徵進行訓練資料的人工收集與標註 (annotation)，可減少分析所需的成本與時間。

關鍵字

情感分析 ； 意見探勘 ； 類別關聯規則 ； 顧客評論

並列摘要

The success and popularity of Web 2.0 applications to encourage a large number of users actively involved of social media; for example: blog, discuss board, social network, etc. The behavior and information sharing of these social media users are valuable assets for corporates; it can help corporates to understand numerous users in order to develop commercial applications. On-line customer evaluation is very important among numerous of social media and users information.It can be valuable information for customers in order to make purchase decision from analyzing customers' evaluation; moreover, it helps retailers and manufacturers to understand their products from the customers' point of view and perspective. Thus helping marketing strategies, product designing, and product improvement.However, the number of post comments from users grow quick fast which makes the user, for example: retailers, manufacturers, customers, must spend a lot of time and effort to find products which interesting and filter out useful information. Analysis manually in order to find out both passive and negative opinion from customers will be difficult and waste time.Therefore, here comes Rule-based Sentiment Analysis technology from this study. Basic on this technology, it collects views of the customer and product prerogative automatically. The cost and time of analysis can be reduced from unnecessary trained labors to do data gather and annotation from each product.

並列關鍵字

Sentiment Analysis ； Opinion Mining ； Class Association Rules ； Consumer Reviews

參考文獻

9. Kobayashi, N., Iida, R., Inui, K., and Matsumotto, Y. (2005). Opinion extraction using a learning-based anaphora resolution technique. In Proceedings of the IJCNLP-05, 173-178.

3. Das, S. and Chen, M., "Yahoo! for Amazon: Sentiment extraction from small talk on the Web," Management Science, Vol. 53, No. 9, 2007, pp. 1375-1388.

4. Ding, X., Liu, B., and Yu, P.S. (2008). A holistic lexicon-based approach to opinion mining. In Proceedings of the WSDM'08, 231-239.

6. Guo, H., Zhu, H., Guo, Z., Zhang, X., and Su, Z. (2009). Product feature categorization with multilevel latent semantic association. In Proceeding of the ACM CIKM'09, 1087-1096.

7. Horrigan, J. B. (2008). Online shopping. Pew Internet & American Life Project, (http://www.pewinternet.org/Reports/2008/Online-Shopping.aspx).