A Study Of Efficient Place Recommendation Techniques In Location-Based Social Networks

The prevalence of Social Networks has changed the way people communicate with others, more and more people like to share and obtain all kinds of information in it. As a result, place recommendation in Location-Based Social Networks has become a fundamental problem. The most general approach for place recommendation is collaborative filtering based on user similarity calculation. However, as it is difficult to find the similar users due to the sparsity problem in location-based social networks, the effectiveness of user-based collaborative recommendation is limited.We think failing to identify the similar places is the most problem in the existing place recommendation. So, in this paper, we first propose a new approach of clustering the similarity places to alleviate the sparsity problem and improve the rate of identifying the similar places, the similarity consists of the semantic of place, represented by tags and category, and the distance between two places, we use three methods when calculate the similarity between places to realize recommendation:(1) clustering by combining single semantic and distance (2)clustering by mixing double semantic and distance (3)clustering by place-based collaborative filtering similarity. Then we analyze users’historical behaviors based on topic modeling, and extract user interests to enhance the traditional collaborative filtering techniques. Finally, we conduct a set of experiments to compare the performance of different methods, the experimental results show both the approaches of clustering place and topic modeling are effective when recommending places.