Design Of Algorithm And Adaptive Index Based On Slide Window For Spatial-Textual Query

With the rapid development of positioning device and internet technology, more and more objects are geo-tagged and associated with textual description. And these objects are called as geo-textual objects. Sometimes, users may want to search geo-textual objects during a period of time in life. But the contents of the queries may vary from user to user. In different condition, query region and query text have different strengths of filtration. At the same time, the query time is limited to the time interval of future, which means that the query has timeliness. However, the current research generally focus on speeding up the rate of filtering query text or region without taking the timeliness of query and different strengths of filtration of query text and region into account in different scenes.So, we proposed an adaptive index based on slid window to deal with continuous query with timeliness in this paper. The slid window is adopted for the timeliness of the query, and a multi-way self-adjust tree is adopted because of the different strengths of filtration between query region and text in different condition. In the adaptive tree, the node of it will adjust itself to a keyword filtering node or a spatial filtering node based on a cost model.Based on this index, a matching algorithm for geo-textual query is proposed.We carry out an experiment based on the real data set and evaluate the effect of parameters of the index through controlling variable method. Then, we compare our index with an index based on Quad-tree plus Inverted File(QIF for short), which demonstrates the index proposed in this paper offer a good performance.