Clustering Of Uncertain Data In Obstacle Space

In recent years,with the continuous understanding of information technology,spatial data clustering has been used in many applications.In the process of data acquisition,it is easy to be affected by the environment or the accuracy of the acquisition instrument,so the data is uncertain and the clustering analysis of uncertain data has higher practical value.In real life,due to some geographical constraints,this paper considers that the constraints of obstacles in real environments are more able to reflect the real space situation,and has higher practical value.Therefore,on the basis of uncertain data clustering,this paper adds obstacle constraints to solve the problem of uncertain data clustering in obstacle space.For the above problems,according to the change of obstacles set,this paper separately solves the clustering problem of static obstacle space and dynamic obstacle space based on attribute level uncertain data.The main research contents are as follows:Firstly,this paper uses probability mass function to represent discrete uncertain data,probability density function to represent continuous uncertain data,and makes use of Kullback-Leibler distance to measure similarity.In order to effectively solve the problem of data clustering in static obstacle space and dynamic obstacle space,Voronoi diagram and grids in computational geometry are introduced to divide the data space as well as an uncertain data clustering algorithm based on Voronoi diagram and grid with obstacle constraints is proposed.Secondly,according to the properties of Voronoi diagram and grid,the clustering rules are proposed.In this paper,the efficiency of the algorithm is improved by analyzing the change of obstacles set.Finally,an uncertain data clustering algorithm is proposed to solve static obstacle space and dynamic obstacle space,in which the dynamic change of obstacle refers to the dynamic increase of obstacle and the dynamic decrease of obstacle.The theory research and experiments show that the proposed RO_UDVBSCAN and G_UDOBSCAN algorithms can be well applied to uncertain data clusteringproblems with obstacle constraints,and they have higher accuracy and efficiency.