地理社交网中基于范围的星型组查询

Star group query based on range in geographical social networks

针对星型组查询只考虑了用户的位置信息和朋友关系,而没有考虑用户的偏好信息,使得推荐质量不高的问题,本文提出一种基于范围的星型组查询算法(m SG),同时考虑用户的社交关系、文本信息和当前位置信息,在给定范围为用户返回k个得分最高的星型组。为了迅速查找中心点用户,提出带倒排表的网格索引结构。为中心点用户更快速地查找朋友,提出了社交关系和文本信息混合索引结构。最后,通过实验表明,利用所提索引结构减少了查询算法的运行时间。

The star group query only considers the user＇ s location information and the friend relationship,without considering the user＇s preference information,which leads to that the quality of recommendation is low. Based on this,a star group query（ m SG）algorithm based on range is proposed,which is based on the user＇ s social relations,text information and current position information. In order to quickly find the center users,a grid with inversion lists index structure is presented. To find a friend quickly for the center users,a hybrid index structure based on social relations and text information is proposed. Finally,experimental results show that the running time of the query algorithm with the proposed index structures is reduced.