摘要
针对传统空间离群检测算法在空间邻居构建上存在算法复杂度过高、人为影响较大等问题,提出一种基于Delaunay三角剖分的空间离群检测算法。该算法通过对空间数据点进行Delaunay三角剖分,建立空间邻域,根据生成的空间邻域,依据SLOF算法定义空间邻域内离群因子。结合某地区矿山具体钻孔数据对该算法进行实验,实验结果表明该算法能有效检测出空间离群点,算法复杂度较低,人为影响较低。
A Delaunay triangulation based spatial outlier detection algorithm is proposed to solve the problems of the traditional spatial outlier detection algorithm, such as the high complexity of the algorithm and the large human influence on the construction of spatial neighbors. In this algorithm, Delaunay triangulation is carried out on the spatial data points to establish the spatial neighbor-hood. Based on the generated spatial neighborhood, outlier factors in the spatial neighborhood are defined according to SLOF algorithm, combined with the specific drilling data of a mine in a certain area. Experimental results show that this algorithm can effectively detect spatial outliers with low complexity and low human influence.
出处
《计算机科学与应用》
2019年第1期1-8,共8页
Computer Science and Application
