A Causality Analysis of Sales by Mining Characteristics of Online Reviews- A Case Study in Movie Social Media

論文翻譯標題: 商品評論特徵探勘模式於銷售量因果關係之研究-以電影社群媒體為例
  • 張 有彤

學生論文: Master's Thesis


Manufacturers have always been interested in the reasons or motivations behind product purchase especially products that involve great investment or have high expectation to succeed Previous studies have demonstrated that online reviews can be considered a form of electronic word-of-mouth and provide interesting information that relate to sales and marketing strategies However due to the exponential growth in the number of reviews companies and customers are finding it increasingly difficult to find desired information The quality of a review is determined by several characteristics e g helpfulness rating writing style reviewers’ emotion etc so it is crucial for manufacturers to find the key characteristics that have a significant effect on sales Most studies in the past only used regression analysis for examining important characteristics of product reviews However correlation does not imply causation meaning that when two variables are correlated one does not necessarily cause the other For the abovementioned reasons this study proposes an approach based on the cause and effect model for comprehensively evaluating characteristics of product reviews To alleviate the issue of information overload opinion mining and retrieval techniques have been used to extract and retrieve complete characteristics from reviews Such characteristics will be indicative of the manufacturer’s sales revenue Results of this study will allow manufacturers understand which indicators to track and manage in order to achieve higher sales
獎項日期2014 6月 25
監督員Sheng-Tun Li (Supervisor)