摘要
LOF(Local Outlier Factor)算法是目前比较实用且效果比较良好的异常点检测算法之一,但是该算法在处理大规模的数据集时,往往会耗费巨大的时间和空间。目前基于网格的异常点检测算法虽然一定情况下降低了算法的时间和空间的耗费,但是时间和空间的耗费依然比较大。对此论文提出一种基于网格山脊点的异常检测算法。该算法先根据数据分布情况划分成空间网格单元,然后计算各个网格山脊点的高度,挑选出网格山脊点低的区域。最后对山脊点低的区域进行LOF算法检测。实验结果表明,相对于目前的基于网格的异常点检测算法,该算法的执行效率显著提高。
The LOF(Local Outlier Factor)algorithm is one of the most practical and effective detection methods. However, the algorithm often takes a lot of time and space when dealing with large-scale data sets. At present,the grid-based outlier detection algorithm reduces the time and space consumption of the algorithm,but the time and space consumption is still relatively large. In this paper,an outlier detection algorithm based on grid ridge is proposed. The algorithm is divided into spatial grid cells according to the data distribution,and then the height of each grid ridge is calculated,and the area with low grid ridge is selected. Finally,the LOF algorithm is detected in the low ridge area. The experimental results show that the efficiency of the algorithm is significantly im. proved compared with the current grid-based outlier detection algorithm.
作者
戴楠
严悍
卓勤政
马玲玲
DAI Nan;YAN Han;ZHUO Qinzheng;MA Lingling(School of Computer Science and Technology,Nanjing University of Science and Technology,Nanjing 210094)
出处
《计算机与数字工程》
2019年第5期1175-1178,共4页
Computer & Digital Engineering
作者简介
戴楠,男,硕士研究生,研究方向:软件工程;严悍,男,博士,副教授,研究方向:信息安全和软件工程;卓勤政,男,硕士研究生,研究方向:软件工程;马玲玲,女,硕士研究生,研究方向:软件工程。
