基于模型的过程工业时间序列异常值检测方法
Model-based outlier detection method for time series of process industry
查看参考文献20篇
文摘
|
时间序列异常点检测是时间序列挖掘研究领域的重要内容和基础工作。为满足过程工业中控制系统大数据量时间序列异常值检测需求,提出了一种计算简单快速的基于边缘化后验比检验的异常值在线检测方法。该方法将基于"偏差"的检测思想与统计学理论相结合,首先利用基于数据的鲁棒建模方法对待检测数据进行拟合并得到拟合残差,然后利用统计学知识分析拟合残差以最终确定数据的异常情况。为了实现算法的在线检测要求,引进了两窗口结构并对其加以改进;通过合理的选取先验分布以及对未知参数进行边缘化处理的方式,有效地减少了算法中的参数个数,提高了算法的可用性。仿真实验可以证明所提出的时间序列异常值在线检测算法具有很好的检测准确性和一定的实用性。 |
其他语种文摘
|
Outlier detection for time series is an important and basic work in time-series data mining research.In order to meet the massive time series outlier detection requirements in control system of process industry,a simple and fast on-line outlier detection method is presented,which is based on marginalized posterior ratio test.The method combines the "deviation" based test idea and statistics theory;row data are fitted by a data-based robust modeling method and fitting residuals are obtained,then the fitting residuals are analyzed with statistics-based method to judge whether the data are outliers or not.To realize on-line detection requirement,double-window structure is introduced and improved;reasonably selecting prior distribution and marginalization processing for unknown parameters effectively decrease the number of parameters and thus improve the availability of the method.Simulation results prove that the proposed method has a superior accuracy and practicability for on-line outlier detection of time series. |
来源
|
仪器仪表学报
,2012,33(9):2080-2087 【核心库】
|
关键词
|
异常值检验
;
时间序列
;
边缘化后验比
;
自回归模型
|
地址
|
1.
中国科学院沈阳自动化研究所, 沈阳, 110016
2.
东北大学信息科学与工程学院, 沈阳, 110819
|
语种
|
中文 |
ISSN
|
0254-3087 |
学科
|
自动化技术、计算机技术 |
基金
|
国家863计划
;
辽宁省自然科学基金
|
文献收藏号
|
CSCD:4642201
|
参考文献 共
20
共1页
|
1.
杜奕.
时间序列挖掘相关算法研究及应用,2007
|
被引
3
次
|
|
|
|
2.
周大镯. 时间序列异常检测.
计算机工程与应用,2008,44(35):145-147
|
被引
8
次
|
|
|
|
3.
Elio L. Parallel algorithms for distancebased and density-based outliers.
International Conference on IEEE,2005:767-776
|
被引
1
次
|
|
|
|
4.
Park Y S. Intrusion detection using noisy training data.
Lecture Notes in Computer Science,2004(3043):547-556
|
被引
1
次
|
|
|
|
5.
刘强. 基于数据和知识的工业过程监视及故障诊断综述.
控制与决策,2010,25(6):801-807, 813
|
被引
29
次
|
|
|
|
6.
许洁. 基于KPCA和MKL-SVM的非线性过程监控与故障诊断.
仪器仪表学报,2010,31(11):2428-2433
|
被引
23
次
|
|
|
|
7.
Fawcett T. Adaptive fraud detection.
Data Mining and Knowledge Discovery,1997,1(3):291-316
|
被引
11
次
|
|
|
|
8.
齐红威. 主曲线异常检测及其在股票市场中的应用.
计算机研究与发展,2005,42(8):1306-1311
|
被引
4
次
|
|
|
|
9.
刘伟新. 主成分回归中异常点的二步诊断法及其医学应用.
现代预防医学,2007,34(13):2423-2425
|
被引
3
次
|
|
|
|
10.
刘芳. 基于模糊自回归隐马尔可夫模型的控制过程异常数据检测.
仪器仪表学报,2010,31(5):984-990
|
被引
8
次
|
|
|
|
11.
Almeida J A S. Improving hierarchical cluster analysis: A new method with outlier detection and automatic clustering.
Chemometrics and Intelligent Laboratory Systems,2007,87(2):208-217
|
被引
11
次
|
|
|
|
12.
Barnet V.
Outlier in statistical data,1994
|
被引
6
次
|
|
|
|
13.
Knorr E M. Finding intentional knowledge of distance-based outliers.
Proceedings of the 25th international Conference on Very Large Data Bases,1999:211-222
|
被引
8
次
|
|
|
|
14.
Knorr E M. Algorithms for mining distancebased outliers in large data sets.
Proceeding of the 24rd International Conference on Very Large Data Bases,1998:392-403
|
被引
1
次
|
|
|
|
15.
Agrawal R. A linear method for deviation detecton in large databases.
Knowledge Discovery and Data Mining Conference,1995:164-169
|
被引
1
次
|
|
|
|
16.
Takeuchi J. A unifying framework for detecting outliers and change points from time series.
IEEE Transactions on Knowledge and Data Engineering,2006,18(4):482-492
|
被引
11
次
|
|
|
|
17.
Alexandridis A. A new algorithm for online structure and parameter adaptation of RBF networks.
Neural Networks,2003,16:1003-1017
|
被引
19
次
|
|
|
|
18.
Alarcon-Aquino V. Anomaly detection in communication networks using wavelet0s.
Communications,IEE,2001,148(6):355-362
|
被引
1
次
|
|
|
|
19.
Gustafsson F. The marginalized likelihood ratio test for detecting abrupt changes.
IEEE Transactions on automatic control,1996,41(1):66-78
|
被引
3
次
|
|
|
|
20.
Bernardo J M.
Bayesian theory,1994
|
被引
15
次
|
|
|
|
|