Abstract
Consumers who shop online are overwhelmed by the unlimited choices and options presented to them. Companies use different recommendation systems (agents), such as collaborative filtering that processes large volumes of data, to provide personalized and relevant recommendations for their customers to alleviate this problem. This paper explores a novel recommendation system that identifies consumers' unique unobserved preferences by utilizing user-generated content (UGC), such as social tags or bookmarks. We propose better metrics and algorithms that can be used as recommendation agents for a successful new product campaign, our proposed method outperforms current popular recommendation algorithms (RA).