大型数据库中数据流异常路径检测仿真
|
摘要
对大型数据库中数据流异常路径进行检测,能够有效保证数据库中数据正确传输。对数据流异常路径的检测,需要选择重叠模块度最大的作为数据流划分结果,给出路径异常度量的依据,完数据流异常路径的检测。传统方法获取数据流可疑异常路径集合,并基于改进距离的异常路径检测算法产生最可能异常路径排序结合,但忽略了得到异常路径度量依据,导致检测精度偏低。提出将数据流的特征属性引入到近似度和重叠模块度计算中的异常路径检测方法,采用K均值聚类对大型数据库数据流分区进行聚类构成均值参考点,利用最小二乘法对生成的数据流均值参考点进行拟合,建立数据流空间邻域,结合混沌粒子群算法自适应惯性权重系数以及对动态学习因子进行改善,提升异常路径检测精确度。实验结果表明,所提方法对于异常路径检测准确,检测精确度较高。
This paper proposes a detection method for abnormal path introducing characteristic attribute of data stream into calculation of degree of approximation and overlap modularity. Our research clustered partition of data stream of large database using K-means cluster to constitute reference point of mean value and used least square method to fit the formed reference point,then built spatial neighborhood of data stream. The research improved the spatial neighborhood integrated with algorithm of chaos particle swarm,coefficient of self-adaption inertia weight and dynamic learning factor and improved detection precision of the abnormal path. Simulation results show that the method detects abnormal path accurately. It has high detection precision.
This paper proposes a detection method for abnormal path introducing characteristic attribute of data stream into calculation of degree of approximation and overlap modularity. Our research clustered partition of data stream of large database using K-means cluster to constitute reference point of mean value and used least square method to fit the formed reference point,then built spatial neighborhood of data stream. The research improved the spatial neighborhood integrated with algorithm of chaos particle swarm,coefficient of self-adaption inertia weight and dynamic learning factor and improved detection precision of the abnormal path. Simulation results show that the method detects abnormal path accurately. It has high detection precision.
引文
[1]叶施仁,等.孤立点预处理和Single-Pass聚类结合的微博话题检测方法[J].计算机应用研究,2016,33(8):2294-2297.
[2]刘露,左万利,彭涛.异质网中基于张量表示的动态离群点检测方法[J].计算机研究与发展,2016,53(8):1729-1739.
[3]任建华,高立明.基于聚类的两段式孤立点检测算法[J].计算机工程与应用,2016,52(20):98-102.
[4]贺鹏程.基于Oracle的数据库性能优化研究[J].电子设计工程,2016,9(24):1-3.
[5]范小刚,朱庆生,万家强.基于K-近邻树的离群检测算法[J].计算机应用研究,2015,32(3):669-673.
[6]冯震,付敬奇,熊南.一种快速的离群点检测方法[J].电子测量与仪器学报,2016,30(11):1726-1734.
[7]杨金鸿,邓廷权.一种基于单簇核PCM的SVDD离群点检测方法[J].电子学报,2017,45(4):813-819.
[8]张玲,刘波.基于残差统计的时间序列加性离群点检测算法研究[J].电子技术应用,2015,41(9):85-87.
[9]洪沙,林佳丽,张月良.基于密度的不确定数据离群点检测研究[J].计算机科学,2015,42(5):230-233.
[10]钱景辉,窦立阳,李荣雨.一种基于多示例学习的局部离群点检测算法[J].信息与控制,2016,45(4):471-478.
[11]陆钊.嵌入式数据库中多维离散数据检测仿真[J].计算机仿真,2017,34(3):400-403.
[2]刘露,左万利,彭涛.异质网中基于张量表示的动态离群点检测方法[J].计算机研究与发展,2016,53(8):1729-1739.
[3]任建华,高立明.基于聚类的两段式孤立点检测算法[J].计算机工程与应用,2016,52(20):98-102.
[4]贺鹏程.基于Oracle的数据库性能优化研究[J].电子设计工程,2016,9(24):1-3.
[5]范小刚,朱庆生,万家强.基于K-近邻树的离群检测算法[J].计算机应用研究,2015,32(3):669-673.
[6]冯震,付敬奇,熊南.一种快速的离群点检测方法[J].电子测量与仪器学报,2016,30(11):1726-1734.
[7]杨金鸿,邓廷权.一种基于单簇核PCM的SVDD离群点检测方法[J].电子学报,2017,45(4):813-819.
[8]张玲,刘波.基于残差统计的时间序列加性离群点检测算法研究[J].电子技术应用,2015,41(9):85-87.
[9]洪沙,林佳丽,张月良.基于密度的不确定数据离群点检测研究[J].计算机科学,2015,42(5):230-233.
[10]钱景辉,窦立阳,李荣雨.一种基于多示例学习的局部离群点检测算法[J].信息与控制,2016,45(4):471-478.
[11]陆钊.嵌入式数据库中多维离散数据检测仿真[J].计算机仿真,2017,34(3):400-403.