Personalized recommendation using implicit interaction information
Nancheng Liu, Qingshan Jiang, Haishan Chen, Beizhan Wang · 2011
Currently, the information in the internet is becoming explosive. In order to help the users searching the items they are interested in, such as, the news, the books, in this paper, we propose an automatic personalized recommendation algorithm by constructing the social graph resting on the users' implicit interaction information. We at first introduce a metric to measure the users' affinity based on their implicit interaction information to construct a social graph, and then categorize the users into different clusters within which they will have similar tastes, finally, we use a personalized recommendation algorithm to recommend the items shared in the same cluster to the users. The experiments on a book data set are performed to demonstrate that our proposed method can well generate the recommendations which users will be interested in with high accuracy and efficiency.