报 告 人：赵少琼 博士
工作单位：美国Carroll University 商学院
赵少琼博士毕业于威斯康辛大学密尔沃基分校。曾多次获得优秀博士奖学金及美国市场营销协会2012年优秀博士生。现任美国Carroll University 商学院助理教授，美国市场营销学学会会员，主要从事市场营销领域电子营销及营销分析、消费者行为、市场调研、文本挖掘等方面的教学科研工作。
Word of mouth has long been recognized to be an influential variable in marketing. With the growth of Internet applications, traditional word of mouth has evolved into the online form in a variety of web-based outlets where individuals spread their perceptions via the written word. These expressions are often in the form of online reviews or assessments of products and services. In this paper we attempt to use features to represent reviews, which contain the sentiments of the consumers, and to predict the overall attitudes of online reviews of the consumers. Further, we want to look at which words are indicative/decision driven of a positive/negative attitude of the consumers, especially we want to identify a set of features which will result in a desired class – positive attitude in our case. Data was collected from a well known website using a WebCrawler type technique and we applied text mining approach for the analysis. The overall results compare favorably with those from standard numeric based quantitative prediction methods. In addition, the text mining methodology and inverse classification help us identify the key features that are related to positive/negative overall attitudes of online users. Identification of key features will be of considerable help to marketers in designing their keyword choices for more effective application of search engine marketing strategies while identification of the negative associated key words will lead to discovery of problematic areas.