Segmentation of Time Series from Nonlinear Dynamical Systems
2011 (English)In: Proceedings of the 18th IFAC World Congress, 2011, 13209-13214 p.Conference paper (Refereed)
Segmentation of time series data is of interest in many applications, as for example in change detection and fault detection. In the area of convex optimization, the sum-of-norms regularization has recently proven useful for segmentation. Proposed formulations handle linear models, like ARX models, but cannot handle nonlinear models. To handle nonlinear dynamics, we propose integrating the sum-of-norms regularization with a least squares support vector machine (LS-SVM) core model. The proposed formulation takes the form of a convex optimization problem with the regularization constant trading off the fit and the number of segments.
Place, publisher, year, edition, pages
2011. 13209-13214 p.
Convex optimization, System identiﬁcation, Support vector machines, Failure detection, Nonlinear systems
IdentifiersURN: urn:nbn:se:liu:diva-95594DOI: 10.3182/20110828-6-IT-1002.01988ISBN: 978-3-902661-93-7OAI: oai:DiVA.org:liu-95594DiVA: diva2:636436
18th IFAC World Congress, Milano, Italy, 28 August-2 September, 2011