A Study of Using Text Mining Techniques to Word of Mouth Analysis

• Main Authors: Lee, Shu-Hui, 李淑惠
• Other Authors: Ker, Sur-Jin
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/10424254576802172574

碩士 === 東吳大學 === 資訊管理學系 === 102 === The opinion has always been important to the consumer based companies develop product strategy. Since the rise of web2.0, more and more consumers through blogs or social networking sites to express their views on the product. Many companies are starting to focus on Internet giant amount of comments information. They hope these comments by the views of consumers to obtain information on the product. However, due to the amount of comments data on the Internet is too great, using the traditional manual approach to analysis is difficult. The purpose of this study hope by automated systems analysis review article on the Internet, automatic collection of different brands on the market reviews data through the system, with features capture, extraction and sentiment analysis techniques to analyze the commentaries. Experimental results show that the system of discrimination in product features and opinions extracting word has a good performance. Through sentiment analysis and product analysis, available on the Internet consumer preferences and the characteristics of the product advantages and disadvantages, and understand the competitive brand analysis results from different brands. These companies analyze data immediately to grasp the market trend, at any time to adjust marketing strategies, and provide the basis for product improvement to the enterprise and analysis of competitors' products.