Change point detection in time series data with random forests

A large class of monitoring problems can be cast as the detection of a change in the parameters of a static or dynamic system, based on the effects of these changes on one or more observed variables. In this paper, the use of random forest models to detect change points in dynamic systems is considered. The approach is based on the embedding of multivariate time series data associated with normal process conditions, followed by the extraction of features from the resulting lagged trajectory matrix. The features are extracted by recasting the data into a binary classification problem, which can be solved with a random forest model. A proximity matrix can be calculated from the model and from this matrix features can be extracted that represent the trajectory of the system in phase space. The results of the study suggest that the random forest approach may afford distinct advantages over a previously proposed linear equivalent, particularly when complex nonlinear systems need to be monitored.

[1]  Nesrin Sarigul-Klijn,et al.  Distance similarity matrix using ensemble of dimensional data reduction techniques: Vibration and aerocoustic case studies , 2009 .

[2]  G. Apostolopoulos,et al.  On-line statistical processing of radiation detector pulse trains with time-varying count rates , 2008 .

[3]  E. Ghysels,et al.  Série Scientifique Scientific Series Monitoring for Disruptions in Financial Markets Monitoring for Disruptions in Financial Markets * , 2022 .

[4]  Monika Ritsch-Marte,et al.  A new method for change-point detection developed for on-line analysis of the heart beat variability during sleep , 2005 .

[5]  Kun Soo Chang,et al.  Applications of chaos and fractals in process systems engineering , 1996 .

[6]  Ratna Babu Chinnam,et al.  General support vector representation machine for one-class classification of non-stationary classes , 2008, Pattern Recognit..

[7]  Christian Boller,et al.  Next generation structural health monitoring and its integration into aircraft design , 2000, Int. J. Syst. Sci..

[8]  Yi-Guang Li,et al.  Gas turbine performance prognostic for condition-based maintenance , 2009 .

[9]  W. Peltier,et al.  Deterministic chaos in the Belousov-Zhabotinsky reaction: Experiments and simulations. , 1993, Chaos.

[10]  Sébastien Borguet,et al.  Coupling principal component analysis and Kalman filtering algorithms for on-line aircraft engine diagnostics , 2009 .

[11]  S. Horvath,et al.  Unsupervised Learning With Random Forest Predictors , 2006 .

[12]  Eunho Ha,et al.  Climatic change and interannual fluctuations in the long‐term record of monthly precipitation for Seoul , 2006 .

[13]  Kenneth A. Loparo,et al.  Automated detection of neonate EEG sleep stages , 2009, Comput. Methods Programs Biomed..

[14]  Caren Marzban,et al.  Using labeled data to evaluate change detectors in a multivariate streaming environment , 2009, Signal Process..

[15]  Emilio Hernández-García,et al.  Ecological thresholds and regime shifts: approaches to identification. , 2009, Trends in ecology & evolution.

[16]  Damián Ginestar,et al.  Analysis of pressure signals using a singular system analysis (SSA) methodology , 2003 .

[17]  Edit Gombay,et al.  Monitoring parameter change in AR(p) time series models , 2009, J. Multivar. Anal..

[18]  Richard J. Field,et al.  Simple models of deterministic chaos in the Belousov-Zhabotinskii reaction , 1991 .

[19]  Mehmet Imregun,et al.  STRUCTURAL DAMAGE DETECTION USING ARTIFICIAL NEURAL NETWORKS AND MEASURED FRF DATA REDUCED VIA PRINCIPAL COMPONENT PROJECTION , 2001 .

[20]  Ingoo Han,et al.  Using change-point detection to support artificial neural networks for interest rates forecasting , 2000 .

[21]  V. Moskvina,et al.  An Algorithm Based on Singular Spectrum Analysis for Change-Point Detection , 2003 .

[22]  Claus Weihs,et al.  Monitoring a deep hole drilling process by nonlinear time series modeling , 2009 .

[23]  D. R. Salgado,et al.  Tool wear detection in turning operations using singular spectrum analysis , 2006 .

[24]  Allan Aasbjerg Nielsen,et al.  Kernel principal component analysis for change detection , 2008, Remote Sensing.

[25]  Hongjoong Kim,et al.  A novel approach to detection of intrusions in computer networks via adaptive sequential and batch-sequential change-point detection methods , 2006, IEEE Transactions on Signal Processing.

[26]  S. Beard,et al.  An Active Diagnostic System for Structural Health Monitoring of Rocket Engines , 2006 .

[27]  Alfredo Vellido,et al.  Advances in clustering and visualization of time series using GTM through time , 2008, Neural Networks.

[28]  Michèle Basseville,et al.  Detection of abrupt changes: theory and application , 1993 .

[29]  Weike Wang,et al.  THE APPLICATION OF SOME NON-LINEAR METHODS IN ROTATING MACHINERY FAULT DIAGNOSIS , 2001 .

[30]  Zhonghua Li,et al.  Adaptive CUSUM control chart with variable sampling intervals , 2009, Comput. Stat. Data Anal..

[31]  Wei Jiang,et al.  Adaptive CUSUM procedures with Markovian mean estimation , 2008, Comput. Stat. Data Anal..

[32]  A. Ray,et al.  An information-theoretic measure for anomaly detection in complex dynamical systems , 2009 .

[33]  D. Carslaw,et al.  Detecting and characterising small changes in urban nitrogen dioxide concentrations , 2007 .

[34]  Tae Yoon Kim,et al.  Variance change point detection via artificial neural networks for data separation , 2005, Neurocomputing.

[35]  Leonhard Held,et al.  Statistical approaches to the monitoring and surveillance of infectious diseases for veterinary public health. , 2009, Preventive veterinary medicine.

[36]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[37]  Yosihiko Ogata,et al.  Statistical model for standard seismicity and detection of anomalies by residual analysis , 1989 .