Model selection approaches for non-linear system identification: a review

Hong, Xia; Mitchell, Richard; Chen, Sheng; Harris, C.J.; Li, K.; Irwin, G.W.

doi:10.1080/00207720802083018

Cited by 202 publications

(148 citation statements)

References 188 publications

Supporting

Mentioning

142

Contrasting

Unclassified

Order By: Relevance

“…A different approach to these sparse kernel modelling methods is the forward selection using the orthogonal-least-squares (OLS) algorithm [8,10], developed in the late 1980s for nonlinear system modelling, which remains highly popular for data-modelling practicians. Since its derivation, many enhanced variants of the OLS forwardselection algorithm have been proposed by incorporating the new developments from machining learning and the approach has extended its application to all the areas of data modelling, including regression, classification and kernel density estimation [9,[11][12][13][14][15][16]7,[25][26][27][28][29]42,33]. This contribution continues this theme, and it presents a unified framework for sparse kernel modelling that include all the three classes of data-modelling applications, namely, regression, classification and PDF estimation.…”

Section: Introductionmentioning

confidence: 90%

“…The derivation of the LOO test error (26) together with the recursive formulas (28) and (29) is detailed in Appendix C. The subset model selection procedure is carried out as follows. At the n-th stage of the selection procedure, a model term is selected among the remaining n to N candidates if the resulting n-term model produces the smallest LOO MSE J n .…”

Section: Article In Pressmentioning

confidence: 99%

See 1 more Smart Citation

Orthogonal-least-squares regression: A unified approach for data modelling

et al. 2009

Self Cite

View full text Add to dashboard Cite

Multiplicative nonnegative quadratic programming a b s t r a c t A unified approach is proposed for data modelling that includes supervised regression and classification applications as well as unsupervised probability density function estimation. The orthogonalleast-squares regression based on the leave-one-out test criteria is formulated within this unified data-modelling framework to construct sparse kernel models that generalise well. Examples from regression, classification and density estimation applications are used to illustrate the effectiveness of this generic data-modelling approach for constructing parsimonious kernel models with excellent generalisation capability.

show abstract

Section: Introductionmentioning

confidence: 90%

Section: Article In Pressmentioning

confidence: 99%

Orthogonal-least-squares regression: A unified approach for data modelling

et al. 2009

Self Cite

View full text Add to dashboard Cite

show abstract

“…The performance of black-box representations used in neurocontrol is highly dependant on the selection of the input space [3,4]. For instance, the choice of neural net inputs may influence computation time, adaptation speed, effects of the curse of dimensionality, understanding of the representation, and model complexity [3,5,6].…”

Section: Introductionmentioning

confidence: 99%

Adaptive structural control using dynamic hyperspace

Laflamme¹

2015

Int. J. CMEM

View full text Add to dashboard Cite

The design of closed-loop structural control systems necessitates a certain level of robustness to cope with system uncertainties. Neurocontrollers, a type of adaptive control system, have been proposed to cope with those uncertainties. However, the performance of neural networks can be substantially influenced by the choice of the input space, or the hyperspace in which the representation lies. For instance, input selection may influence computation time, adaptation speed, effects of the curse of dimensionality, understanding of the representation, and model complexity. Input space selection is often overlooked in literature, and inputs are traditionally determined offline for an optimized performance of the neurocontroller. Such offline input selection is often unrealistic to conduct in the case of civil structures. In this paper, a novel method for automating the input selection process for neural networks is presented. The method is purposefully designed for online input selection during adaptive identification and control of nonlinear systems. Input selection is conducted online and sequentially, while the excitation is occurring. The algorithm designed for the adaptive input space assumes local quasi-stationarity of the time series, and embeds local maps sequentially in a delay vector using the embedding theorem. The input space of the representation is subsequently updated. The performance of the proposed dynamic input selection method is demonstrated through simulating semi-active control of an existing structure located in Boston, MA, U.S.A. Simulation results show the substantial performance of the proposed algorithm over traditional fixed-inputs strategies.

show abstract

“…In order to achieve this goal, the analyst should combine the information arising from experimental data with the background knowledge on the functioning of a particular dynamical system. The mathematical techniques used in the field of the system identification serve a threefold objective, namely: they can be employed for identifying a set of unknown parameters of a nonlinear mathematical model by using an experimental dataset; they can be used for the determination of the model order corresponding to a given system; and they allow for obtaining finite-dimensional linear models of a general mechanical system starting from an interrelated sequence of input-output data [26,27]. The methods of system identification can be applied to deterministic systems, as well as to stochastic processes [28,29].…”

Section: Introductionmentioning

confidence: 99%

System Identification Algorithm for Computing the Modal Parameters of Linear Mechanical Systems

Pappalardo

Guida

2018

Machines

View full text Add to dashboard Cite

Abstract:The goal of this investigation is to construct a computational procedure for identifying the modal parameters of linear mechanical systems. The methodology employed in the paper is based on the Eigensystem Realization Algorithm implemented in conjunction with the Observer/Kalman Filter Identification method (ERA/OKID). This method represents an effective and efficient system identification numerical procedure based on the time domain. The algorithm developed in this work is tested by means of numerical experiments on a full-car vehicle model. To this end, the modal parameters necessary for the design of active and semi-active suspension systems are obtained for the vehicle system considered as an illustrative example. In order to analyze the performance of the methodology developed in this investigation, the system identification numerical procedure was tested considering two case studies, namely a full state measurement and an incomplete state measurement. As expected, the numerical results found for the identified dynamical model showed a good agreement with the modal parameters of the mechanical system model. Furthermore, numerical results demonstrated that the proposed method has good performance considering a scenario in which the signal-to-noise ratio of the input and output measurements is relatively high. The method developed in this paper can be effectively used for solving important engineering problems such as the design of control systems for road vehicles.

show abstract

Model selection approaches for non-linear system identification: a review

Cited by 202 publications

References 188 publications

Orthogonal-least-squares regression: A unified approach for data modelling

Orthogonal-least-squares regression: A unified approach for data modelling

Adaptive structural control using dynamic hyperspace

System Identification Algorithm for Computing the Modal Parameters of Linear Mechanical Systems

Contact Info

Product

Resources

About