Detecting changes of opinion from customer reviews

• Main Authors: Zhi-Han Ke, 柯志翰
• Other Authors: Li Chen Cheng
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/20138340440014285005

碩士 === 東吳大學 === 資訊管理學系 === 100 === With the flourishing of the World Wide Web, the online customer review process is becoming more and more useful and important as an information resource for people. As a result, opinion mining research for analysis of opinion data on the web has recently become a popular topic. Most previous studies have used Pointwise Mutual Information (PMI) to predict the opinion (positive or negative). This study first proposes a method which combines the associative classification methodology with the overall rating to discover the relation of the features. By the way, the proposed framework discovered the changes of users’ opinion that can identify the characteristics of different generations. The tested products included the tablet ipad, ipad2 and Camera LX3, LX5, S90 and S95. Experimental results demonstrate the effectiveness of the proposed approach. The data set contains actual WOM information from which to study the dynamic patterns of users' changing opinions. The summarized results can help consumers and marketing managers to make decision.