The Community for Technology Leaders
RSS Icon
Issue No.09 - September (2009 vol.31)
pp: 1537-1551
Christopher K.I. Williams , University of Edinburgh, Edinburgh
Neil McIntosh , University of Edinburgh, Edinburgh
Condition monitoring often involves the analysis of systems with hidden factors that switch between different modes of operation in some way. Given a sequence of observations, the task is to infer the filtering distribution of the switch setting at each time step. In this paper, we present factorial switching linear dynamical systems as a general framework for handling such problems. We show how domain knowledge and learning can be successfully combined in this framework, and introduce a new factor (the “X-factor”) for dealing with unmodeled variation. We demonstrate the flexibility of this type of model by applying it to the problem of monitoring the condition of a premature baby receiving intensive care. The state of health of a baby cannot be observed directly, but different underlying factors are associated with particular patterns of physiological measurements and artifacts. We have explicit knowledge of common factors and use the X-factor to model novel patterns which are clinically significant but have unknown cause. Experimental results are given which show the developed methods to be effective on typical intensive care unit monitoring data.
Condition monitoring, switching linear dynamical system, switching Kalman filter, novelty detection, intensive care.
Christopher K.I. Williams, Neil McIntosh, "Factorial Switching Linear Dynamical Systems Applied to Physiological Condition Monitoring", IEEE Transactions on Pattern Analysis & Machine Intelligence, vol.31, no. 9, pp. 1537-1551, September 2009, doi:10.1109/TPAMI.2008.191
[1] D. Alspach and H. Sorenson, “Nonlinear Bayesian Estimation Using Gaussian Sum Approximation,” IEEE Trans. Automatic Control, vol. 17, pp. 439-447, 1972.
[2] R. Shumway and D. Stoffer, “Dynamic Linear Models with Switching,” J. Am. Statistical Assoc., vol. 86, pp. 763-769, 1991.
[3] N. de Freitas, R. Dearden, F. Hutter, R. Morales-Menedez, J. Mutch, and D. Poole, “Diagnosis by a Waiter and a Mars Explorer,” Proc. IEEE, vol. 92, no. 3, 2004.
[4] R. Morales-Menedez, N. de Freitas, and D. Poole, “Real-Time Monitoring of Complex Industrial Processes with Particle Filters,” Advances in Neural Information Processing Systems 15, S. Becker, S.Thrun, and K. Obermayer, eds., MIT Press, 2002.
[5] U. Lerner, R. Parr, D. Koller, and G. Biswas, “Bayesian Fault Detection and Diagnosis in Dynamic systems,” Proc. 17th Nat'l Conf. Artificial Intelligence, pp. 531-537, 2000.
[6] V. Pavlović, J. Rehg, and J. MacCormick, “Learning Switching Linear Models of Human Motion,” Advances in Neural Information Processing Systems 13, T. Leen, T. Dietterich, and V. Tresp, eds., MIT Press, 2000.
[7] Y. Li, T. Wang, and H.-Y. Shum, “Motion Texture: A Two-Level Statistical Model for Character Motion Synthesis,” Proc. ACM SIGGRAPH '02, pp. 465-472, 2002.
[8] M. Azzouzi and I. Nabney, “Modelling Financial Time Series with Switching State Space Models,” Proc. IEEE/IAFE Conf. Computational Intelligence for Financial Eng., pp. 240-249, 1999.
[9] A. Smith and M. West, “Monitoring Renal Transplants: An Application of the Multiprocess Kalman Filter,” Biometrics, vol. 39, pp. 867-878, 1983.
[10] J. Droppo and A. Acero, “Noise Robust Speech Recognition with a Switching Linear Dynamic Model,” Proc. IEEE Int'l Conf. Acoustics, Speech, and Signal Processing, 2004.
[11] J. Ma and L. Deng, “A Mixed Level Switching Dynamic System for Continuous Speech Recognition,” Computer Speech and Language, vol. 18, pp. 49-65, 2004.
[12] A. Cemgil, H. Kappen, and D. Barber, “A Generative Model for Music Transcription,” IEEE Trans. Speech and Audio Processing, vol. 14, no. 2, pp. 679-694, 2006.
[13] C. Williams, J. Quinn, and N. McIntosh, “Factorial Switching Kalman Filters for Condition Monitoring in Neonatal Intensive Care,” Advances in Neural Information Processing Systems 18, Y.Weiss, B. Schölkopf, and J. Platt, eds., MIT Press, 2006.
[14] J. Quinn and C. Williams, “Known Unknowns: Novelty Detection in Condition Monitoring,” Proc. Third Iberian Conf. Pattern Recognition and Image Analysis, J. Martí, J.-M. Benedí, A.M.Mendonça, and J. Serrat, eds., 2007.
[15] J. Quinn, “Neonatal Condition Monitoring Demonstration Code,”, 2008.
[16] K. Tsien, “Dynamic Bayesian Networks: Representation, Inference and Learning,” PhD dissertation, Univ. of California, Berkeley, 2002.
[17] Z. Ghahramani and G. Hinton, “Parameter Estimation for Linear Dynamical Systems,” technical report, Dept. of Computer Science, Univ. of Toronto, 1996.
[18] Z. Ghahramani and G. Hinton, “Variational Learning for Switching State-Space Models,” Neural Computation, vol. 12, no. 4, pp.963-996, 1998.
[19] J. Candy, Model-Based Signal Processing. Wiley-IEEE Press, 2005.
[20] P.C. Woodland, “Hidden Markov Models Using Vector Linear Prediction and Discriminative Output Distributions,” Proc. IEEE Int'l Conf. Acoustics, Speech, and Signal Processing, vol. I, pp. 509-512, 1992.
[21] Z. Ghahramani and M. Jordan, “Factorial Hidden Markov Models,” Machine Learning, vol. 29, pp. 245-273, 1997.
[22] M. West and J. Harrison, Bayesian Forecasting and Dynamic Models. Springer, 1999.
[23] J. Quinn, “Bayesian Condition Monitoring in Neonatal Intensive Care,” PhD dissertation, Univ. of Edinburgh, http://www.era.lib. 1645, 2007.
[24] M. Markou and S. Singh, “Novelty Detection: A Review—Part 1: Statistical Approaches,” Signal Processing, vol. 83, pp. 2481-2497, 2003.
[25] J. Ma and S. Perkins, “Online Novelty Detection on Temporal Sequences,” Proc. Ninth ACM SIGKDD Int'l Conf. Knowledge Discovery and Data Mining, pp. 613-618, 2003.
[26] P. Smyth, “Markov Monitoring with Unknown States,” IEEE J. Selected Areas in Comm., vol. 12, no. 9, pp. 1600-1612, 1994.
[27] U. Lerner and R. Parr, “Inference in Hybrid Networks: Theoretical Limits and Practical Algorithms,” Proc. 17th Ann. Conf. Uncertainty in Artificial Intelligence, pp. 310-318, 2001.
[28] K. Murphy, “Switching Kalman Filters,” technical report, Univ. of California, Berkeley, 1998.
[29] K. Murphy and S. Russell, “Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks,” Sequential Monte Carlo in Practice, A. Doucet, N. de Freitas, and N. Gordon, eds., Springer-Verlag, 2001.
[30] J. Hunter and N. McIntosh, “Knowledge-Based Event Detection in Complex Time Series Data,” Proc. Joint European Conf. Artificial Intelligence in Medicine and Medical Decision Making, W. Horn, Y.Shahar, G. Lindberg, S. Andreassen, and J. Wyatt, eds., 1999.
[31] S. Miksch, W. Horn, C. Popow, and F. Paky, “Utilizing Temporal Data Abstraction for Data Validation and Therapy Planning for Artificially Ventilated Newborn Infants,” Artificial Intelligence in Medicine, vol. 8, no. 6, pp. 543-576, 1996.
[32] I.J. Haimowitz, P.P. Le, and I.S. Kohane, “Clinical Monitoring Using Regression Based Trend Templates,” Artificial Intelligence in Medicine, vol. 7, no. 6, pp. 473-496, 1995.
[33] M. Imhoff, M. Bauer, U. Gather, and D. Löhlein, “Statistical Pattern Detection in Univariate Time Series of Intensive Care In-Line Monitoring Data,” Intensive Care Medicine, vol. 24, pp. 1305-1314, 1998.
[34] S. Hoare and P. Beatty, “Automatic Artifact Identification in Anaesthesia Patient Record Keeping: A Comparison of Techniques,” Medical Eng. and Physics, vol. 22, pp. 547-553, 2000.
[35] S. Charbonnier, G. Becq, and G. Biot, “On-Line Segmentation Algorithm for Continuously Monitored Data in Intensive Care Units,” IEEE Trans. Biomedical Eng., vol. 51, no. 3, pp. 484-492, 2004.
[36] R. Kennedy, “A Modified Trigg's Tracking Variable as an 'Advisory' Alarm during Anaesthesia,” J. Clinical Monitoring and Computing, vol. 12, pp. 197-204, 1995.
[37] R. Shumway and D. Stoffer, Time Series Analysis and Its Applications. Springer-Verlag, 2000.
[38] A. Spengler, “Neonatal Baby Monitoring,” master's thesis, School of Informatics, Univ. of Edinburgh, 2003.
[39] J. Hunter, “TSNet—A Distributed Architecture for Time Series Analysis,” Proc. Intelligent Data Analysis in bioMedicine and Pharmacology, pp. 85-92, 2006.
[40] D. Barber and B. Mesot, “A Novel Gaussian Sum Smoother for Approximate Inference in Switching Linear Dynamical Systems,” Advances in Neural Information Processing Systems 18, Y. Weiss, B.Schölkopf, and J. Platt, eds., MIT Press, 2006.
18 ms
(Ver 2.0)

Marketing Automation Platform Marketing Automation Tool