Catching Preference Drift with Initiators in Social Network

User's preference drift over time gets it difficult to make accurate recommendation. A recommender system ignoring the fact always recommends similar items which were loved previously by users while users' preference have changed and new items appear. It has been proven empirically that traditional algorithms handling the concept drift problem which simply takes time into account is not appropriate, so a model considering both the static and dynamic preference well is the key to catch users' preference drift. We put forward an original model which explores the behavior of influential people, the initiators who initiate trends in social network, to handle this problem. Compared to traditional collaborative filtering approaches and time weighted approaches, empirical study on lastfm dataset has shown that our model improves the accuracy of the recommendation.