中文摘要：

摘要为了帮助移动用户探索新位置，协助商家发现潜在顾客，提高移动用户访问POI的质量，本文提出了一种将地理位置、时间和类别相结合的多关联度POI个性化推荐方法——GeoTempCate，来对移动用户进行个性化POI推荐．该方法探讨了POI的地理位置关联关系、访问时间关联关系以及POI的类别关联关系．通过分析用户的历史访问POI数据，利用3种关联关系来预测移动用户对其未访问POI的关联评分，进而对移动用户进行个性化POI推荐．最后使用准确率和召回率来评价推荐的POI，实验结果表明本文提出的方法，无论是准确率还是召回率，都得到了较高的精确度．

英文摘要：

In order to help mobile users explore new POIs, assist businesses to discover potential customers, and improve the quality of mobile user＇s access to POIs, we, in this paper, propose a personalized POI recommendation approach for mobile users, called GeoTempCate. This approach discusses such correlations as geographical location correlations, temporal correlations and categorical correlations of POI. With the analysis of the historical user access POI data, the approach is able to use the three related correlations to predict the associated rating of POI that has not been visited by the mobile user so as to perform the personalized POI recommendation to the mobile user. Finally, we use precision rate and recall rate to evaluate the personalized POI recommendation. The experimental results show that GeoTempCate gains a higher accuracy in both precision rate and recall rate.