中文摘要：

兴趣点推荐是基于位置的社会网络的重要研究内容之一。传统的兴趣点推荐算法或者应用基本的协同过滤方法,或者在基本的协同过滤算法中引入空间特征,而没有充分发掘时序特征对推荐算法的重要性。为了进一步提高兴趣点推荐算法的性能,提出了一种面向时序特征的兴趣点推荐算法。给出了基本的基于用户的协同过滤方法,分别描述了时间特征和空间特征的作用,并给出了相应的模型表示方法;将时间特征和空间特征进行融合,提出了一种联合推荐算法。实验表明,提出的算法与其他相关算法相比,准确率和召回率显著提高,因此更适合兴趣点的推荐服务。

英文摘要：

Point of interest recommendation is a critical issue in location based social networks. Traditional recommender algorithms used either naive collaborative filter algorithms, or space feature based collaborative filter algorithms. However, these algorithms neglected the importance of time feature in point of interest recommendation. In order to improve the performance of algorithms in point of interest recommendation, this paper proposed a time feature based point of interest recommender algorithm in location based social networks. Firstly, it described the naive user based collaborative filter algorithm. Secondly, it analyzed the importance of time and space features separately, and proposed corresponding models. Finally, it fused the time and space features, and proposed a unified recommender algorithm. The experiments show that, the proposed algorithm has better precision and recall compared with related works, and thus can be used in real point of interest recommender services.