Exploiting web reviews for generating customer service surveys

Traditional customer satisfaction analysis relies on the work of designing, distributing, collecting and analyzing surveys. Surveys that are designed by humans may be subjective, and it is hard to know what service aspects are the most important for customers. To address this issue, this paper proposes a method of automatically generating service surveys through mining Web reviews. Candidate service aspects are extracted using simple extraction rules. Then we rank candidate service aspects in terms of their weights generated by combining co-occurrence method and linear regression method together. Experimental results demonstrate the effectiveness of the proposed method.