User interest modeling based on large item-set group

In order to solve the issue of data sparsity and user´s multiple interests which exist in the practical application, we proposes a kind of user interest modeling method based on the large item-set group. The mapping relationship between the items and the user evaluations is transformed into the one between the item properties and the user evaluations to solve the problem of data sparsity; the large item-set group is used to solve the problem of user´s multiple interests. This modeling method has been verified by collecting data from www.douban.com. The results showed that our recommended method based on the large item-set group can effectively reduce the difference level between the target users and the recommended knowledge, comparing with other current recommended methods.