Abstract:

An increasing number of location based social networking services emerge, and location recommendation systems are one of its main applications. Currently, more and more scholars study location recommendation algorithms, which are committed to improve the prediction of users’ location interest degree without considering the time factor. In fact whether or not people will visit a place is closely related to timing. We therefore propose a probabilistic model, which takes the three factors, which are users’ interest degree inthe place, timing and the popularity of the place, into account under the unified framework. We test the algorithm on a real data set extracted from Foursquare, and the results show that our method can achieve better recommendation compared with the classical approaches, and it can enhance users’ experience as well.

Key words: location based social network;location recommendation;Bayesian principle