Research on Distance Join Selectivity Estimation for Spatial Database

Spatial join query is widely used in geographic information system.Accurate and fast selectivity estimation for spatial join is profitable for optimizing the join query.Incorporating exist selectivity estimation techniques,authors firstly propose a distance join selectivity estimation method based on MBR buffer and histogram.The method based on the MBR buffer can be used to obtain satisfying estimation results fast by computing only several statistics of histogram.Since these methods produce larger results than true value,a new selectivity estimation method for distance join based on line segment distribution characteristic is proposed to improve the estimation based on MBR buffer and histogram.Line segment distribution law is according to statistical characteristic of realistic datasets.Without traversing the line segment datasets,the method based on the line segment buffer provides accurate results enjoying less relative error.The experimental results show that the methods are more efficient for spatial distance query selectivity estimation,and especially more applicable to realistic datasets.