Power of bosom friends, POI recommendation by learning preference of close friends and similar users

Abstract

With the emergence of social networks, mining interesting information from the social media datasets becomes a popular research direction. Previous researches on social networks, such as POI (point of interest) recommendation, usually ignore the social tie strength between users. If we can further consider the closeness between friends in the analysis, it is possible to improve the results. Therefore, in this paper, we focus on analyzing the social tie strength between users in the location-based social network. The proposed method analyzes the movement of users and the interaction between them by the spatial-temporal data. Furthermore, the social relationship structure is also taken into consideration for the calculation of the social tie strength. Finally, the location list for POI recommendation will be constructed accordingly. Experimental results show that the proposed method significantly outperforms the competitor on both precision and recall.