WorksEngineering
Bayesian Learning of Uncertain Linear System Behaviours
Abstract
We present a Bayesian approach to learning uncertain behavioural representations of linear dynamical systems from noisy input-output data. In behavioural systems theory, the finite-horizon behaviour of a linear system is represented by a subspace. We quantify uncertainty in this subspace through probability distributions on the Grassmann manifold. This leads to a geometric formulation of Bayesian inference over the Grassmannian. We derive the posterior distribution under Gaussian measurement noise and discuss how uncertainty in the inferred behaviour can be propagated to trajectory predictions, with a view towards data-driven control.
Introduction
Behavioural systems theory [1] characterizes a linear time-invariant (LTI) system through its subspace of admissible trajectories. A representation of this subspace can be constructed from sufficiently informative input–output data [2], without first identifying a parametric model. Over finite horizons, the restricted behaviour is represented by the column space of a suitably constructed data matrix. This perspective supports methods for system identification [3], [4], [5] and data-enabled control [6], [7], [8], with demonstrated practical effectiveness [8], [9].
In practice, noise corrupts both the offline training data and the online measurements used for feedback and prediction, and can severely degrade controller performance if its effects are not accounted for [10]. Quantifying and compensating for these effects is therefore a fundamental challenge in both system identification and data-driven control. This has motivated formulations that extend beyond the deterministic, noise-free setting [8], including regularized optimization [6], [10], robust formulations [11], and distributionally robust formulations [7], which can improve robustness but often introduce hyperparameters that require tuning. Other strategies preprocess data before constructing the behavioural representation, for example through low-rank approximation [12], including truncated singular value decomposition (SVD) [13]. Maximum a posteriori estimation offers another approach to trajectory inference under uncertainty [14].
We propose an uncertainty quantification framework for restricted behaviours of linear dynamical systems. The central idea is to represent uncertainty about the restricted behaviour by a probability distribution on the Grassmann manifold and learn this distribution from data using Bayesian inference. Motivated by [15], [16], this formulation respects the invariance of the subspace to the choice of basis and quantifies uncertainty arising from noisy observations. It also allows uncertainty to be propagated to finite-horizon trajectory predictions using tools from geometric probability and computation on manifolds [17], [18], [19], [20].
The statistical analysis of estimation error and model uncertainty has long been central to system identification [21]. Uncertainty quantification for identified LTI models has been developed for state-space [22], [23], [24], frequency-response [25], [26], and ARMA representations [27]. For predictors constructed from behavioural representations, uncertainty in the predicted outputs is quantified in [28], [29], and a deterministic analysis of prediction error is given in [30]. Related work has also investigated uncertainty quantification and geometric distances for restricted behaviours [15], [16].
Closely related research provides complementary perspectives on probabilistic descriptions of dynamical systems. Faulwasser et al. [31] use polynomial chaos expansions (PCE) to obtain a deterministic representation of stochastic dynamics in a higher-dimensional coefficient space. Sasfi et al. [32] model finite-length trajectories as Gaussian random vectors, characterizing trajectory variability through their covariance structure. Finally, Preciado [33] develops a measure-theoretic formulation based on probability measures supported on admissible trajectories, leading to convex formulations of optimal control problems. Our focus is Bayesian inference over restricted behaviours themselves, with uncertainty in the inferred behaviour propagated to trajectory predictions.
Preliminaries
Let and be the sets of symmetric positive semidefinite and positive definite matrices, respectively. For a matrix , we write for its transpose, for its Moore–Penrose pseudoinverse, for its determinant, and for its th largest singular value. The identity matrix is . We use the abbreviations
where is symmetric positive definite. The Gaussian distribution with mean and covariance is written , with density
We will make use of two types of smooth manifolds in this paper: the compact Stiefel manifold and the Grassmann manifold [34] (see also [35], [36]). Subspaces are denoted by bold uppercase letters and their orthonormal basis matrices by the corresponding uppercase letters. For integers $n > 0 and , with , the Grassmann manifold is the set of all -dimensional linear subspaces of
The compact Stiefel manifold is the set of all matrices in with orthonormal columns
The compact Stiefel manifold is an embedded submanifold of , and the Grassmann manifold is its quotient under orthogonal changes of basis [34, 36], where is the orthogonal group of matrices. Thus, a subspace is represented by any whose columns span . Two representatives and describe the same subspace if and only if for some .
We can also identify with its orthogonal projector , which is independent of the representative [37, 38]. For with , the chordal distance is defined by
where is the Frobenius norm. Equivalently,
where , with , are the principal angles between the subspaces and , and satisfy [39, 40].
The compact Grassmann manifold carries a uniform probability measure [37]. Throughout the paper, densities on are taken with respect to . For a symmetric matrix , the matrix Bingham distribution on has density [37]
where is the confluent hypergeometric function of matrix argument [41], which can be evaluated numerically using the algorithms in [42].
Problem Setup
We consider a discrete-time LTI system
where , , and denote the state, input, and output, respectively. The system matrices , , , and are unknown, and the state is unobserved.
For integers and , define the stacked input and output vectors
The stacked vector
is an admissible length- input–output trajectory of (2) if there exists a state sequence satisfying (2) at each time . The restricted behaviour of (2) consists of all admissible input–output trajectories of (2). The restricted behaviour of (2) consists of all admissible input–output trajectories of (2). The restricted behaviour is a linear subspace of , where and . We denote its dimension by . For an observable realization (2), for , where is the observability index of [2].
The stacked vector
is an admissible length- input–output trajectory of (2) if there exists a state sequence satisfying (2) at each time . The restricted behaviour of (2) consists of all admissible input–output trajectories of (2). The restricted behaviour of (2) consists of all admissible input–output trajectories of (2). The restricted behaviour is a linear subspace of , where and . We denote its dimension by . For an observable realization (2), for , where is the observability index of [2].
Our objective is to infer the restricted behaviour of (2) from noisy observations and provide a principled quantification of uncertainty in the inferred subspace. Inputs and outputs are assumed to be measured with additive Gaussian noise, giving an errors-in-variables model [43]. Specifically, we consider the probabilistic model
where is independent of the input-output trajectory and such that .
Subspace Representation from Data
We first recall how a restricted behaviour can be represented in the absence of noise. Consider a measured trajectory
For a positive integer , define the depth- Hankel matrix , by
If is controllable and is persistently exciting of order , the fundamental lemma [2, 44] gives
Thus, every admissible trajectory can be written as
for some . The construction extends to multiple records by concatenating their Hankel matrices under a collective excitation condition [44]. For length- input-output trajectories, the corresponding noise-free data matrix represents whenever it has rank .
An orthonormal representation follows from the compact SVD [8],
where satisfies , and . Thus, the restricted behaviour is a point on [15], [16], and represented by any of its orthonormal basis matrices. For noisy measurements, the rank of a Hankel matrix may be larger than the dimension of underlying noise-free restricted behaviour. In this case, standard practice is to truncate the singular value decomposition by retaining dominant singular directions to obtain an estimate of the subspace [8].
Probabilistic Beliefs over Behaviours
The core idea in this paper is to model finite-horizon behaviours as elements of the Grassmann manifold and to learn the belief
where is a data matrix of observed trajectories
Uncertainty about the restricted behaviour is modelled by a probability distribution3 on . Thus, our framework replaces a deterministic restricted behaviour with a probability measure over admissible behaviours. A distribution can subsequently be propagated into probabilistic behavioural predictive control and simulation problems, providing a principled mechanism for representing epistemic uncertainty in data-driven predictions.
Inference
Let denote the unknown restricted behaviour and let be any representative, so that and . We use the observation model
where all are independent, and , with being a hyperparameter (usually estimated using expectation maximization) and being a known noise covariance. This model (7) generates noisy samples
from . Noise-free samples satisfy with a distribution that does not depend on the specific representative but only the rotation-invariant projection . The latent vector can be thought of as a prior on the initial state of the system and serves as a hyperparameter. The choice of is not critical (as a prior its influence will be forgotten as the dataset grows), so we take throughout for simplicity. The variance of must be isotropic however, if instead with a fixed , then the noise-free trajectories depend on the representation rather than the subspace .
By Bayes’ rule, the posterior density of with respect to is
with
Next, we examine the likelihood , prior , and normalizing constant (or evidence) .
The likelihood: We first introduce the sample covariance matrix and aggregated precision, defined as
respectively. The next result characterizes the likelihood.
Proposition 1 (Likelihood): Consider the observation model (7) and the data matrix defined in (6). Assume that and are known. Then, , i.e., for every ,
where and are defined in (10) and
Proof: Fix a representative of . By (7), the are i.i.d. with , . Since is an orthogonal projector of rank , and [45]. Hence
which is (11), and depends on only through .
Proposition 1 shows that the log-likelihood is an affine function of and that the data enter only through and, as a function of , the likelihood is proportional to a matrix Bingham density. The likelihood can also be naturally expressed using projectors. Some algebraic manipulation yields
where . is known as the “misfit” and the terms are a measure of distance between the data and the subspace [46].
Remark 1: (Maximum Likelihood Estimation) The Maximum Likelihood Estimate (MLE)
provides some insights about the likelihood. By Ky Fan’s maximum principle [47] , where is a matrix of the dominant Eigenvectors of . This equivalent to the PCA estimate of [45] and the column space of the truncated SVD of the data matrix used in subspace identification [3] and in subspace predictive control [48]. The subspace is unique if and only if , where are the eigenvalues of . This condition holds almost surely under Gaussian noise whenever , i.e., we have enough data relative to the rank of the system.
Selection of the prior
We use priors of the form
in the matrix Bingham family (1). We consider two specific choices of the prior. The choice gives the uninformative uniform prior , that is, the invariant distribution . Alternatively, let be a reference subspace with projector obtained for example by a SVD of an independent data matrix . Then the informative prior
is a decreasing function of the chordal distance to and has mode . The hyperparameter thus quantifies trust in and (14) reduces to the uniform prior for .
The evidence
For both choices of prior, the evidence in (9) cannot be written in closed-form but can be expressed in terms of hypergeometric functions. With the uniform prior,
by Proposition 1 and (1). With the prior (14),
Both functions can be evaluated with the algorithm of [42] or by Monte Carlo integration with exact samples from [49].
The posterior distribution
With the likelihood and prior in hand, we can derive expressions for the posterior.
Theorem 1 (Posterior). Consider the observation model (7), the data matrix , and and as in (10). Assume that and are known and that the prior is for a symmetric . Then the posterior is distributed according to
In particular, the uniform prior gives and the prior (14) gives .
Proof. By Proposition 1 and (1), equals times a factor that does not depend on . Normalizing with respect to by (1) gives (15).
Theorem 1 shows that the matrix Bingham family is conjugate to the likelihood of Proposition 1. Conjugacy follows because the log-likelihood is affine in , the sufficient statistic of the exponential family on . The data enter through the additive update . Since grows linearly with , a fixed prior contribution becomes negligible relative to the data term as .
Remark 2: (The MAP Estimate) The maximum a posteriori (MAP) estimate is
By Ky Fan’s principle, it can be computed as , where the columns of are eigenvectors associated with the largest eigenvalues of . For the prior , the MAP estimate coincides with the MLE when and converges to as . The concentration of the posterior around the MAP estimate is governed by the gaps between the largest eigenvalues of and the remaining eigenvalues.
The condition (4) for noise-free data has a natural statistical counterpart. One may require the distribution of the underlying trajectories to span the restricted behaviour, i.e.,
Under i.i.d. sampling with finite second moments and isotropic measurement noise, the MLE then converges almost surely to as . The MAP estimate has the same limit for a fixed Bingham prior. The observation model (7) satisfies this condition automatically.
Prediction
Let and let
be a new trajectory, where is independent of and distributed as (7). We permute and partition the trajectory as , where collects the initial inputs , the initial outputs and the future inputs , and collects the future outputs we wish to predict. We also partition a representative and the projector,
in the same way. The following theorem characterizes the predictive distribution.
Theorem 2 (Predictive distribution). Consider the observation model (7), the data matrix , and the new trajectory . Assume that and are known and let be our belief about . Then:
i) For any fixed , with
ii) the predictive density is
where . Replacing by in (18) omits the additional information about supplied by . The approximation is accurate when is approximately constant over the region where concentrates, as may occur for sufficiently large .
ii) Since and are conditionally independent given , and . The claim follows from the law of total probability and Theorem 1.
Theorem 2 expresses the predictive density as a continuous mixture of Gaussian densities weighted by the posterior . The resulting density need not be Gaussian and may be multimodal when substantial uncertainty about the restricted behaviour remains. The integral over the manifold can be evaluated by quadrature in low dimensions or approximated using Riemannian Monte Carlo methods in higher dimensions.
When the posterior is concentrated near a point estimate , such as the MLE or MAP estimate, one may use the approximation
where the right-hand side is the Gaussian density characterized in Theorem 2. For fixed , the noise-free limit gives
If has full column rank, then and, for compatible , the limiting prediction is
This recovers the noise-free predictor of [50, 48], evaluated at . For , the full column rank condition holds if , where is the system lag.
Numerical Case Study
Consider a first-order autonomous system
where , with additive measurement noise on the output according to and . We choose this system because its restricted behaviour for is a line in , allowing direct visualization of uncertainty in the inferred subspace. We generate trajectories of length . We rewrite the system as
A basis for the noise-free trajectory subspace is deduced from this representation as
yielding the noise-free subspace as a line through the origin angled at .
Inference: Assuming an uninformative uniform prior, we compute the posterior density over a subset of candidate subspaces . Figure 1 shows the posterior density with true subspace and MAP estimates. Figure 2 illustrates that the angular error scales as we gather more data. Figure 3 demonstrates the evolution of the subspace angle MAP estimate as a function of the signal-to-noise ratio (SNR), .

Figure 1. Posterior density with true subspace and MAP estimates , and for , and samples, respectively.

Figure 2. Principal angle error between MAP subspace estimate and true subspace as a function of for with corresponding trendline. The error decreases as . Each data point represents the average over multiple independent noise realizations to obtain a more representative estimate of the underlying trend.

Figure 3. MAP subspace estimate and corresponding 95% confidence interval as a function of the signal-to-noise ratio, using samples. The posterior converges to a Dirac measure at the true subspace as
Prediction: Next, we perform prediction as in § 3-B. We focus on the MAP conditioned distribution for computational feasibility as evaluating the full predictive posterior will require specialized sampling algorithms which we defer to an extension.
Here and, writing with , the noise-free predictor is . Interestingly, the posterior density of is bounded below by a positive constant and , so has no posterior mean. Since is increasing on , the posterior quantiles of are of the posterior quantiles of . Figure 4 compares the output distribution obtained using the true subspace with the predictive distribution obtained by propagating subspace uncertainty according to Theorem 1. Finally, Figure 5 illustrates the dependence of the MAP-based output prediction on the SNR.

Figure 4. Posterior densities for with true noisy and posterior with and , using a fixed MAP subspace estimate rather than the full posterior distribution. The true and predicted distributions are close but offset, illustrating how errors in the estimated subspace propagate to the output prediction and introduce systematic bias. For and , the posterior median of is 0.768 and its 95% credible interval is [0.695, 0.840].

Figure 5. MAP output estimate and corresponding 95% confidence interval as a function of signal-to-noise ratio, calculated using fixed MAP subspace estimate with . This predictive distribution is conditioned on a single MAP subspace estimate and neglects uncertainty in the underlying subspace, producing overconfident predictions. The full predictive distribution (18) should suffer less from this issue. However, the true prediction does reliably lie in the 95% confidence interval indicating an acceptable level of approximation error.
Summary: At a high level, the numerical example indicates that: 1) the framework is consistent in the sense that the posterior distribution and predictive distribution approach that of the data generating process (almost surely) as or and 2) the MAP conditioned distribution is overconfident but still provides a useful prediction and degree of uncertainty quantification.
Conclusions and Future Work
The paper introduced a Bayesian framework for learning restricted behaviours from noisy input–output data. Uncertainty about the restricted behaviour is represented by a probability distribution on the Grassmann manifold. Under the Gaussian observation model considered, we established conjugacy of the matrix Bingham family and derived predictive distributions that account for uncertainty in the inferred behaviour. A numerical case study illustrated the inference and prediction procedures.
Future work includes extending the observation model beyond isotropic noise, developing scalable algorithms for probabilistic prediction in larger systems, and exploring non-parametric representations of uncertainty about the restricted behaviour. Applications to model fusion and model predictive control could combine information from first-principles models with operational data while accounting for uncertainty in the inferred behaviour.
References
- [1]J. C. Willems, “From time series to linear system—Part I. Finite dimensional linear time invariant systems,” Automatica, vol. 22, no. 5, pp. 561–580, Sep. 1986.DOI
- [2]J. C. Willems, P. Rapisarda, I. Markovsky, and B. L. De Moor, “A note on persistency of excitation,” Syst. Control Lett., vol. 54, no. 4, pp. 325–329, Apr. 2005.
- [3]P. van Overschee and B. L. R. De Moor, Subspace Identification for Linear Systems: Theory - Implementation - Applications. Boston, MA, USA: Kluwer Acad. Publ, 1996.
- [4]I. Markovsky, J. C. Willems, P. Rapisarda, and B. L. De Moor, “Algorithms for deterministic balanced subspace identification,” Automatica, vol. 41, no. 5, pp. 755–766, May 2005.DOI
- [5]M. Verhaegen and V. Verdult, Filtering and System Identification: A Least Squares Approach, 1st ed. Cambridge, U.K.: Cambridge University Press, Apr. 2007.
- [6]J. Coulson, J. Lygeros, and F. Dörfler, “Data-enabled predictive control: In the shallows of the DeePC,” in 2019 18th European Control Conference (ECC). Naples, Italy: IEEE, Jun. 2019, pp. 307–312.arxiv.org/abs/1811.05890
- [7]J. Coulson, J. Lygeros, and F. Dörfler, “Distributionally robust chance constrained data-enabled predictive control,” IEEE Trans. Automat. Contr., vol. 67, no. 7, pp. 3289–3304, Jul. 2022.DOI
- [8]I. Markovsky and F. Dörfler, “Behavioral systems theory in data-driven analysis, signal processing, and control,” Annu. Rev. Control, vol. 52, pp. 42–64, 2021.DOI
- [9]I. Markovsky, L. Huang, and F. Dörfler, “Data-Driven Control Based on the Behavioral Approach: From Theory to Applications in Power Systems,” IEEE Control Syst., vol. 43, no. 5, pp. 28–68, Oct. 2023.DOI
- [10]F. Dörfler, J. Coulson, and I. Markovsky, “Bridging direct and indirect data-driven control formulations via regularizations and relaxations,” IEEE Trans. Automat. Contr., vol. 68, no. 2, pp. 883–897, Feb. 2023.
- [11]L. Huang, J. Zhen, J. Lygeros, and F. Dörfler, “Robust data-enabled predictive control: Tractable formulations and performance guarantees,” IEEE Trans. Automat. Contr., vol. 68, no. 5, pp. 3163–3170, May 2023.
- [12]I. Markovsky, Low Rank Approximation: Algorithms, Implementation, Applications. London, U.K.: Springer, 2012.
- [13]I. Markovsky and F. Dörfler, “Data-driven dynamic interpolation and approximation,” Automatica, vol. 135, Jan. 2022.DOI
- [14]M. Yin, A. Ianelli, S. A. Nazari, and M. A. Müller, “A unified Bayesian framework for data-driven smoothing, prediction, and control,” Apr. 2026, https://arxiv.org/abs/2512.01475.
- [15]A. Padoan, J. Coulson, H. J. Van Waarde, D. Lygeros, and F. Dörfler, “Behavioral uncertainty quantification for data-driven control,” in 2022 IEEE 61st Conference on Decision and Control (CDC). Cancun, Mexico: IEEE, Dec. 2022, pp. 4726–4731.
- [16]A. Padoan and J. Coulson, “Distances between finite-horizon linear behaviors,” IEEE Control Syst. Lett., vol. 9, pp. 1772–1777, 2025.arxiv.org/abs/2503.22849
- [17]M. Girolami and B. Calderhead, “Riemann Manifold Langevin and Hamiltonian Monte Carlo Methods,” Journal of the Royal Statistical Society Series B: Statistical Methodology, vol. 73, no. 2, pp. 123–214, Mar. 2011.DOI
- [18]M.-N. Tran, D. H. Nguyen, and D. Nguyen, “Variational Bayes on manifolds,” Stat Comput, vol. 31, no. 6, p. 71, Nov. 2021.arxiv.org/abs/1908.03097
- [19]L. Lin, V. Rao, and D. Dunson, “Bayesian nonparametric inference on the Stiefel manifold,” Statist. Sinica, vol. 27, no. 2, pp. 535–553, 2017.arxiv.org/abs/1311.0907
- [20]P. Turaga, A. Veeraraghavan, and R. Chellappa, “Statistical analysis on Stiefel and Grassmann manifolds with applications in computer vision,” in 2008 IEEE Conference on Computer Vision and Pattern Recognition. Anchorage, AK, USA: IEEE, Jun. 2008, pp. 1–8.DOI
- [21]L. Ljung, System Identification: Theory for the User, 2nd ed. Upper Saddle River, NJ, USA: Prentice Hall, 1999.
- [22]A. Chiuso and G. Picci, “Numerical conditioning and asymptotic variance of subspace estimates,” Automatica, vol. 40, no. 4, pp. 677–683, Apr. 2004.DOI
- [23]A. Tsiamis, I. Ziemann, N. Matni, and G. J. Pappas, “Statistical Learning Theory for Control: A Finite-Sample Perspective,” IEEE Control Syst., vol. 43, no. 6, pp. 67–97, Dec. 2023.arxiv.org/abs/2209.05423
- [24]A. R. Mesquita, “Robust and Bayesian subspace identification,” IEEE Trans. Automat. Contr., vol. 70, no. 2, pp. 1395–1401, Feb. 2025.DOI
- [25]S. Damjanovic, J. W. Simpson-Porco, and K. van Heusden, “Distributionally robust frequency response estimation using non-steady-state frequency-domain data,” Oct. 2026, https://arxiv.org/abs/2610.09395.
- [26]Y. Pedari, F. Kaviani, and H. R. Ossareh, “Error bounds for direct data-driven frequency-response estimation,” International Journal of Control, vol. 99, no. 4, pp. 901–911, Apr. 2026.DOI
- [27]T. Söderström and P. Stoica, System Identification. New York, NY, USA: Prentice Hall, 1989.
- [28]M. Yin, A. Ianelli, and R. S. Smith, “Data-driven prediction with stochastic data: Confidence regions and minimum mean-squared error estimates,” in 2022 European Control Conference (ECC). London, United Kingdom: IEEE, Jul. 2022, pp. 853–858.DOI
- [29]F. Kaviani, I. Markovsky, and H. R. Ossareh, “Uncertainty quantification of data-driven output predictors in the output error setting,” IEEE Trans. Automat. Contr., vol. 70, no. 11, pp. 7588–7595, Nov. 2025.arxiv.org/abs/2404.15098
- [30]D. Jin and J. Coulson, “On the sensitivity of the subspace predictor to behavioral perturbations,” IEEE Control Syst. Lett., vol. 10, pp. 1609–1614, 2026.
- [31]T. Faulwasser, R. Ou, G. Pan, P. Schmitz, and K. Worthmann, “Behavioral theory for stochastic systems? A data-driven journey from Willems to Wiener and back again,” Annu. Rev. Control, vol. 55, pp. 92–117, 2023.
- [32]A. Sasfi, I. Markovskyy, A. Padoan, and F. Dörfler, “Gaussian behaviors: representations and data-driven control,” in 2025 IEEE 64th Conference on Decision and Control (CDC). Rio de Janeiro, Brazil: IEEE, Dec. 2025, pp. 1993–1998.arxiv.org/abs/2504.15838
- [33]V. M. Preciado, “A measure-theoretic formulation of behavioral systems,” May 2026, http://arxiv.org/abs/2605.01558.
- [34]P.-A. Absil, R. Mahony, and R. Sepulchre, “Riemannian geometry of Grassmann manifolds with a view on algorithmic computation,” Acta Appl. Math., vol. 80, no. 2, pp. 199–220, Jan. 2004.DOI
- [35]N. Boumal, An Introduction to Optimization on Smooth Manifolds, 1st ed. Cambridge, U.K.: Cambridge University Press, Mar. 2023.DOI
- [36]P.-A. Absil, R. Mahony, and R. Sepulchre, Optimization Algorithms on Matrix Manifolds. Princeton, NJ, USA: Princeton Univ. Press, Dec. 2008.DOI
- [37]Y. Chikuse, Statistics on Special Manifolds. New York, NY, USA: Springer, 2003.
- [38]T. Bendokat, R. Zimmermann, and P.-A. Absil, “A Grassmann manifold handbook: Basic geometry and computational aspects,” Adv Comput Math, vol. 50, no. 1, p. 6, Feb. 2024.DOI
- [39]J. H. Conway, R. H. Hardin, and N. J. A. Sloane, “Packing lines, planes, etc.: Packings in Grassmannian spaces,” Exp. Math., vol. 5, no. 2, pp. 139–159, Jan. 1996.DOI
- [40]A. Edelman, T. A. Arias, and S. T. Smith, “The geometry of algorithms with orthogonality constraints,” SIAM J. Matrix Anal. Appl., vol. 20, no. 2, pp. 303–353, Jun. 1998.arxiv.org/abs/physics/9806030
- [41]R. J. Muirhead, Aspects of Multivariate Statistical Theory. Hoboken, NJ, USA: Wiley-Interscience, 2005.
- [42]P. Koev and A. Edelman, “The efficient evaluation of the hypergeometric function of a matrix argument,” Math. Comp., vol. 75, no. 254, pp. 833–846, Jan. 2006.arxiv.org/abs/math/0505344
- [43]T. Söderström, Errors-in-Variables Methods in System Identification. Cham, Switzerland: Springer, 2018.
- [44]H. J. Van Waarde, C. De Persis, M. K. Camlibel, and P. Tesi, “Willems’ fundamental lemma for state-space systems and its extension to multiple datasets,” IEEE Control Syst. Lett., vol. 4, no. 3, pp. 602–607, Jul. 2020.arxiv.org/abs/2002.01023
- [45]M. E. Tipping and C. M. Bishop, “Probabilistic principal component analysis,” J. R. Statist. Soc. Ser. B (Statist. Methodol.), vol. 61, no. 3, pp. 611–622, Sep. 1999.DOI
- [46]I. Markovsky, J. C. Willems, S. Van Huffel, and B. De Moor, Exact and Approximate Modeling of Linear Systems: A Behavioral Approach. Philadelphia, PA, USA: SIAM, Jan. 2006.
- [47]K. Fan, “On a theorem of Weyl concerning eigenvalues of linear transformations I,” Proc. Natl. Acad. Sci. U.S.A., vol. 35, no. 11, pp. 652–655, Nov. 1949.
- [48]W. Favoreel, B. D. Moor, and M. Gevers, “SPC: Subspace predictive control,” IFAC Proc. Vol., vol. 32, no. 2, pp. 4004–4009, Jul. 1999.DOI
- [49]F. Mezzadri, “How to generate random matrices from the classical compact groups,” Feb. 2007, https://arxiv.org/abs/math-ph/0609050.
- [50]I. Markovsky and P. Rapisarda, “Data-driven simulation and control,” Int. J. Control, vol. 81, no. 12, pp. 1946–1959, Dec. 2008.DOI