Top-K Frequent Spatial-Temporal Words Query Based on R-Tree

The top-k query has always been a research hotspot in spatial data queries. With the development of social software, users can mark their geographical location when publishing information, and new data analysis problems come behind, such as the user wants to know some of the hottest words in a certain area in the recent period of time. In this paper, we define this kind of query problem as kHSW (Top-k Heat Spatial-Temporal Words), which returns k words with the highest heat among all posts in the region by given a space-time region. In order to solve this problem, we propose an index-plus-algorithm approach based on R-tree spatial index, in which RIL list is built for R-tree's leaf nodes to improve index efficiency. In addition, we optimize the classic algorithm TA and proposed OTA algorithm to fit the index structure that we use in this paper. Finally, a large number of comparative experiments verify the high efficiency of the soolution.