Abstract
User reviews on Web sites offer the product evaluation by consumers' view point which is different from providers' view point. However, thier contents are the mixture of good and bad ones because reviews with less credibility by the stealth marketing might lurk in them. Then, the purpose of this study is to support credibility judgment of user review, focusing on reviews by the stealth marketing. The text mining technique is used for finding reviews by the stealth marketing. It has the possibility to detects the tendency of unnatural reviews from the correlation between time and the appearance of key words. As a result, a new material to judge the user review credibility is offered.
