A method of generating customer's profile without history for providing recommendation to new customers in E-commerce

Abstract

One of the advantages in E-commerce is that the long tail marketing strategy can be employed. By this, customers can get recommendations about the items, which are rare and specialized to their own tastes. In order to provide this long tail based recommendation service, the service provider needs to have knowledge about the each user's preference and the similarity among the items which have their own peculiar. If the customer's purchasing transaction history is provided, his/her preference can be inferred through data mining techniques. But if a customer is new and the purchasing history is empty, it is hard to extract the collect profile for the customer. In this paper, a method of defining the customer's profile through collective intelligence is proposed. This method can generate profile even if the customer's personal history does not exist. Therefore a proper recommendation can be provided to newcomers in the service. © 2012 Springer Science+Business Media Dordrecht.