关键词: 海量数据, 活动轨迹, 海量活动轨迹的网格索引（GATH）, 近似查询

Abstract: Given a sequence of query locations, each associated with a set of key activities, an activity trajectory similarity query returns k trajectories that cover the query activities and yield the shortest minimum match distance. Since GAT （Grid index for Activity Trajectories） is not for big data, it introduces a new structure GATH （GAT on Hadoop） to solve the problem of similarity search on massive activity trajectories. Moreover, GATH uses grid index for space pruning and inverted index for keyword pruning. The experimental results demonstrate that GATH is more efficient for both index building and data pruning than GAT.

Key words: massive data, active trajectories, Grid index for Activity Trajectories on Hadoop（GATH）, similarity search