Pith. sign in

REVIEW 4 major objections 4 minor 37 references

Efficient Bayesian Inference for Discretely Observed Continuous Time Markov Chains

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read A pseudo-likelihood coupling the observed transition matrix to a biorthogonal spectral form of the generator makes Bayesian inference for discretely observed continuous-time Markov chains both scalable and asymptotically justified.

desk verdict Genuinely useful computational idea for CTMC inference, but the theoretical claims overstate what is proved and the implemented sampler targets an approximation of the pseudo-posterior; still worth serious refereeing. read the letter →

arxiv 2507.16756 v1 pith:P7CKZ7GK submitted 2025-07-22 stat.ME stat.CO

classification stat.MEstat.CO MSC 62M0562F1560J27
keywords continuous-timeMarkovchaingeneratormatrixpseudo-likelihoodBayesianinferenceBernstein-vonMisestheorembiorthogonalspectraldecompositionembeddabilityGibbssampling
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes a Bayesian method for continuous-time Markov chains observed only at discrete time points, where the exact likelihood is intractable. Instead of integrating over unobserved holding times, it couples the transition matrix P with the generator L through a pseudo-likelihood that penalizes differences between P and the spectral form of L. The resulting Gibbs sampler has cost nearly invariant to the number of observations and scales to state spaces with tens to hundreds of states. The authors prove a Bernstein–von Mises theorem for P and posterior consistency for the spectral parameters of L, and demonstrate the method on simulations and a metastable diffusion.

What carries the argument

The load-bearing object is the biorthogonal spectral decomposition $L = \sum_{k=1}^m \lambda_k \phi_k \psi_k^\top$ with $\psi_j^\top \phi_k = \delta_{jk}$, under which $\exp(\Delta L) = \sum_k \exp(\lambda_k \Delta) \phi_k \psi_k^\top$. The pseudo-likelihood combines the observed-transition log-likelihood with a Frobenius penalty $\|P - \sum_k \exp(\lambda_k \Delta) \phi_k \psi_k^\top\|_F^2$ scaled by $\nu > 0$. This couples the empirical transition matrix to the generator, and the conditional updates for P, the eigenvalues, and the eigenvectors take conjugate-friendly forms: Dirichlet for the rows of P, and truncated normal or multivariate normal for the spectral parameters. Biorthogonality is encouraged by a shrinkage prior toward the Stiefel manifold and truncations that keep off-diagonal entries of L nonnegative.

What would settle it

Simulate a three-state CTMC whose generator is defective (a single Jordan block, so no biorthogonal decomposition exists), observe it at regular intervals, and run the proposed Gibbs sampler: if the posterior contracts to a generator that misrepresents the true transition probabilities or the spectral penalty residual remains large, the biorthogonal assumption is doing the work.

Watch

Extended reading notes

Core claim

The central claim is that the intractable likelihood of a discretely observed CTMC can be replaced by a pseudo-likelihood that is both tractable and asymptotically justified: the log-likelihood of the observed transitions plus a Frobenius penalty tying P to the biorthogonal spectral representation $L = \sum_{k=1}^m \lambda_k \phi_k \psi_k^\top$. Under this pseudo-likelihood, the posterior of P is asymptotically normal at the parametric rate, and the spectral parameters of L converge at $\sqrt{n}$, with the rate explicitly showing that closely spaced eigenvalues inflate the uncertainty in eigenvectors. The payoff is a Gibbs sampler that avoids latent path integration, obeys embeddability, and returns an embeddable generator with coherent posterior uncertainty.

Load-bearing premise

The data-generating generator must be diagonalizable with a biorthogonal spectral decomposition, and the observations must be regularly spaced; if the true generator is defective or the observation times are irregular, the stated guarantees do not follow.

Editorial extensions

If this is right

  • Bayesian inference for discretely observed CTMCs no longer requires sampling latent paths, so computation is nearly invariant to the number of observations.
  • The posterior distribution of the transition matrix P concentrates at the parametric rate, giving calibrated uncertainty quantification.
  • The spectral parameters of the generator converge at $\sqrt{n}$, and the convergence rate reveals that closely spaced eigenvalues make eigenvectors harder to estimate.
  • The method returns an embeddable generator, avoiding the failure mode where an empirical transition matrix has no valid matrix-logarithm.
  • The approach scales to medium-to-high dimensional state spaces, demonstrated on a 30-state metastable diffusion example.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the asymptotic theorems fix the regularization strength $\nu$, the finite-sample bias-variance tradeoff is not addressed; one testable extension is to let $\nu$ grow slowly with n and examine the posterior coverage of P.
  • The biorthogonal assumption rules out defective generators; for near-defective systems, the posterior may concentrate on a diagonalizable approximation, and the Frobenius residual could serve as a diagnostic for that mismatch.
  • The same pseudo-likelihood coupling could be extended to irregularly spaced observations by modeling P as a function of the time lag, rather than fixing a single $\Delta$.
  • When eigenvalues are close, the inflated eigenvector uncertainty provides a principled way to flag non-identifiability in the embedding problem.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper proposes a scalable Bayesian approach for inference on discretely observed continuous-time Markov chains. It defines a pseudo-likelihood that combines the observed transition log-likelihood with a Frobenius penalty coupling the transition matrix P to a biorthogonal spectral decomposition of the generator L (Eq. 6). A Gibbs sampler (Algorithm 1) alternates between sampling rows of P from Dirichlet distributions based on transition counts and sampling the spectral parameters from truncated Gaussian full conditionals. The paper claims a Bernstein-von Mises theorem for P and posterior consistency for the spectral parameters of L, and supports the method with simulations and a metastable diffusion application. The central promise is near-cost-invariant inference with coherent posterior uncertainty, bypassing latent path integration.

Significance. If the method worked as claimed, it would be a useful contribution: it avoids latent path sampling, respects embeddability within the assumed spectral class, and appears computationally attractive for state spaces of tens to hundreds of states. The paper also provides code and an R package, and it compares against exact latent-path methods. However, the gap between the theoretical results and the actual sampler is substantial, and the current manuscript does not establish the advertised coherent posterior uncertainty. The biorthogonal assumption and the restriction to real spectra are also underexplored. The core algorithmic idea is promising, but the theoretical framing and implementation claims need significant reworking.

major comments (4)
  1. [Section 3.1 and Algorithm 1, line 5] The implemented sampler does not target the pseudo-posterior defined by Eq. (6). The full conditional of P under (6) is non-conjugate; Section 3.1 argues that it converges to Dir(α + c_p) as n → ∞ or ν → 0, and Algorithm 1 line 5 samples P directly from this limiting Dirichlet, with only a parenthetical suggestion to use it as an MH proposal. No Metropolis-Hastings acceptance step is reported, so the stationary distribution of the composite update is not the pseudo-posterior (6). The simulations use ν = 10^4 and n as small as 10^2, so the penalty term is far from negligible in the regime where the approximation is used. This undermines the claim that the sampler yields coherent posterior uncertainty for L in the finite-sample settings presented.
  2. [Section 4.2, Theorem 2] Theorem 2 is a frequentist delta-method for the spectral decomposition of the plug-in empirical generator L̂ = Δ^{-1} log P̂, not a posterior consistency result for the spectral parameters under the pseudo-posterior (6) or under Algorithm 1's sampling distribution. The theorem's statement concerns √n(λ̂_k − λ_k^0), √n(φ̂_k − φ_k^0), and √n(ψ̂_k − ψ_k^0) for the empirical estimator, not the posterior samples produced by the sampler. The abstract's claim of 'posterior consistency for the spectral parameters' is therefore not supported by the displayed result. Additionally, the proof's formula for the left eigenvector, √n(ψ̂_k − ψ_k^0) → ψ_k^0 Q R_k, does not match the theorem statement's √n(ψ̂_k − ψ_k^0) → R_k^T Q^T ψ_k^0; these differ and the dimensions should be checked.
  3. [Section 3.2 and Eq. (8)] The method as implemented is restricted to generators with real eigenvalues. The prior in Eq. (8) constrains the eigenvalues to 1 = Λ_1 > Λ_2 ≥ ... ≥ Λ_m > 0, and the eigenvectors are modeled as real Gaussian vectors in Eq. (9). However, the biorthogonal representation L = Σ λ_k φ_k ψ_k^T in Section 2.2 allows, in general, complex eigenvalues and eigenvectors for non-reversible CTMCs, and Section 2.2 even orders eigenvalues by Re(λ_k). For a non-reversible generator with a complex conjugate pair, the proposed sampler cannot run. This scope restriction is not stated as a modeling assumption in the abstract or introduction, and the simulation studies only use symmetric generators, so the method's applicability to general CTMCs is overstated.
  4. [Section 4.1 and Theorem 1] The Bernstein-von Mises theorem in Theorem 1 concerns the posterior P(P | x_n) under the Dirichlet-multinomial likelihood, not the pseudo-posterior of Eq. (6). Even if one accepts the algorithm's replacement of the P-conditional with Dir(α + c_p), Theorem 1 does not account for the spectral penalty in (6); hence it cannot justify the two-stage procedure used in Algorithm 1. What is needed is a statement about the joint pseudo-posterior, or an explicit analysis of the approximate sampler's stationary distribution, neither of which is provided.
minor comments (4)
  1. [Table 2] In the m = 4, n = 10^4 row for the 'blandt' method, the reported Frobenius norm is 0.16 with a standard deviation of 0.79; a standard deviation larger than the mean by a factor of five seems implausible and is likely a typographical error.
  2. [Theorem 2 statement and proof] The covariance matrix of Q is stated as 2 I^{-1}(P_0) without derivation; the factor 2 and the precise form of the Fisher information for a Markov chain with row-wise multinomial counts should be made explicit, since the asymptotic covariance of the empirical transition matrix is not simply the inverse Fisher information of a single transition.
  3. [Section 2.2] The notation uses m for both the state-space size and the number of spectral components; this is acceptable only if it is clearly stated that the spectral truncation uses all m components. A short clarification would help.
  4. [Section 3.1] The phrase 'converges in total variation of the conditional posterior' is imprecise: the convergence in Eq. (7) should be stated as convergence of the conditional posterior distribution of P(p, ·) to the specified Dirichlet distribution, and under which norm (total variation) it holds.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper's equations and fitted parameters are not shown to reduce to their own inputs by construction; the main weaknesses are theoretical gaps and a sampler mismatch, which are not circular.

full rationale

The pseudo-likelihood in (6) does couple P to the spectral parameters through the Frobenius penalty, but the implemented Gibbs sampler explicitly breaks that coupling: Section 3.1 states that the full conditional of P under (6) is non-conjugate and that it converges to Dir(alpha + c_p) as n -> inf or nu -> 0, and Algorithm 1 line 5 samples P directly from Dir(alpha + c_p) without a Metropolis-Hastings acceptance step. That means the sampler's stationary distribution is not the joint pseudo-posterior defined by (6), and the spectral penalty does not feed back into P. This is a validity gap, not a circular reduction: no equation is equivalent to its input by construction, and the spectral parameters are genuinely updated conditional on P through the penalty. Similarly, Theorem 2 is a frequentist delta-method for the plug-in spectral estimator of the empirical transition matrix, not a posterior consistency theorem for the spectral parameters sampled by Algorithm 1, so the abstract's claim of posterior consistency for the spectral parameters overstates the theory; again this is a mismatch, not circularity. The biorthogonal spectral representation is stated as an explicit modeling assumption in Section 2.2, and the prior relaxation in (9) follows Duan et al. (2020) and Matuk et al. (2022), with an optional Stiefel projection attributed to Astfalck et al. (2024); these are published or externally checkable methods, and no self-citation is invoked as an unverified uniqueness theorem to forbid alternatives. The empirical comparisons against the external benchmarks Bladt and Sorensen (2005) and Riva-Palacio et al. (2023) provide independent support. Overall, the paper's derivation chain is not circular in the sense defined by the requested patterns, even though the stated posterior guarantees do not cover the algorithm as implemented.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The method's central claims rest primarily on the biorthogonal representation of L, the Perron-Frobenius ordering of eigenvalues, and the approximate Dirichlet update for P. The hyperparameters (ν, σ_φ^2, σ_ψ^2, σ_c^2, α) are chosen by hand and jointly control the strength of regularization; they are free parameters in the inference. No new entities are introduced.

free parameters (4)
  • ν (regularization strength) = 10^4
    Controls the strength of the Frobenius penalty pulling P toward its spectral form; chosen by hand, no sensitivity analysis. Section 2.4 and Section 5.1.
  • σ_φ^2, σ_ψ^2 (eigenvector prior variances) = 10^{-1}
    Regularize the scale of the right and left eigenvectors to aid identifiability; chosen by hand. Section 5.1.
  • σ_c^2 (biorthogonality shrinkage) = 10^{-5}
    Shrinks left and right eigenvectors toward biorthogonality; product ν σ_c^2 = 10^{-1} sets the effective shrinkage. Chosen by hand.
  • α (Dirichlet prior parameter for rows of P) = (1,...,1)
    Uniform prior on each row of the transition matrix; a standard noninformative choice.
assumptions (5)
  • domain assumption The generator L admits a biorthogonal spectral decomposition L = Σ λ_k φ_k ψ_k^T with φ_k^T ψ_l = δ_{kl}.
    Section 2.2 states 'Assuming the spectral representation (2) exists.' This is the key restriction that makes the exponential and logarithm computable and unique; defective generators are excluded.
  • domain assumption The chain is irreducible and aperiodic, so Perron-Frobenius gives Λ_1=1 and |Λ_i|<1 for i≥2.
    Section 2.3 uses this to conclude rank(L)=m-1 and to justify the eigenvalue ordering and prior range in (8).
  • domain assumption For the BvM theorem, the true transition matrix has full support (A1) and the prior is a truncated Dirichlet (A2).
    Section 4.1 introduces (A1)-(A2) to guarantee recurrence, geometric ergodicity, and identifiability of P.
  • standard math Connault's (2014) weakly dependent Bernstein-von Mises theorem is valid and applicable after verifying conditions (C1)-(C7).
    Section 4.1 and Appendix B use Connault's theorem as the backbone of Theorem 1; the verification is sketched in the appendix.
  • domain assumption Discretizing the Smoluchowski diffusion into interval states induces a CTMC with a generator that approximates the coarse-grained dynamics.
    Section 5.2 adopts this 'under standard regularity conditions' to apply the method to the metastable diffusion example.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Bayesian Inference for Discretely Observed Continuous Time Markov Chains." pith.science (2026). https://pith.science/paper/P7CKZ7GK

@misc{pith2026250716756,
  author       = {Pith},
  title        = {Pith review of: Efficient Bayesian Inference for Discretely Observed Continuous Time Markov Chains},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P7CKZ7GK}},
  note         = {Machine review of arXiv:2507.16756}
}
read the original abstract

Inference for continuous-time Markov chains (CTMCs) becomes challenging when the process is only observed at discrete time points. The exact likelihood is intractable, and existing methods often struggle even in medium-dimensional state-spaces. We propose a scalable Bayesian framework for CTMC inference based on a pseudo-likelihood that bypasses the need for the full intractable likelihood. Our approach jointly estimates the probability transition matrix and a biorthogonal spectral decomposition of the generator, enabling an efficient Gibbs sampling procedure that obeys embeddability. Existing methods typically integrate out the unobserved transitions, which becomes computationally burdensome as the number of data or dimensions increase. The computational cost of our method is near-invariant in the number of data and scales well to medium-high dimensions. We justify our pseudo-likelihood approach by establishing theoretical guarantees, including a Bernstein-von Mises theorem for the probability transition matrix and posterior consistency for the spectral parameters of the generator. Through simulation and applications, we showcase the flexibility and robustness of our approach, offering a tractable and scalable approach to Bayesian inference for CTMCs.

Figures

Figures reproduced from arXiv: 2507.16756 by the authors.

Figure 1
Figure 1. Run-times per 1000 MCMC samples. The model [PITH_FULL_IMAGE:figures/full_fig_p014_1.png] view at source ↗
Figure 2
Figure 2. Left, the three-well potential U(x) governing the dynamics in the Smoluchowski equation. Center, reference transition matrix estimated from a long simulation of 5 × 107 time steps. Right, empirical transition matrix Pˆ estimated from a subsampled trajectory of 10,000 observations. 0.00 0.18 0.36 5 2.5 0 −2.5 −5 0.00 0.18 0.36 −7.8 −2.2 3.4 0.00 0.03 0.06 5 2.5 0 −2.5 −5 −5 −2.5 0 2.5 5 0.00 0.01 0.02 −5 −2.5 0 2.5 5… view at source ↗
Figure 3
Figure 3. Top row, posterior means of the transition matrix [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Left, expected difference E[P − P˜ | xn]. Center, difference between E[P | xn] and the reference matrix. Right, difference between E[P˜ | xn] and the reference matrix. the expected difference E[P−P˜ | xn] (left), the deviation between the reference matrix and E[P | xn]…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 34 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 'mid.sentence := #2 '...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Asmussen, S. (2003). Applied probability and queues, vol. 2. Springer

  4. [4]

    , Sen, D

    Astfalck, L. , Sen, D. , Patra, S. , Cripps, E. & Dunson, D. (2024). Posterior projection for inference in constrained spaces. arXiv preprint , 1812.05741

  5. [5]

    Billingsley, P. (1995). Probability and Measure. Wiley Series in Probability and Mathematical Statistics. New York: John Wiley & Sons, 3rd ed

  6. [6]

    , Bouchard-C \^o t \'e , A

    Biron-Lattes, M. , Bouchard-C \^o t \'e , A. & Campbell, T. (2023). Pseudo-marginal inference for CTMCs on infinite spaces via monotonic likelihood approximations. Journal of Computational and Graphical Statistics 32, 513--527

  7. [7]

    & S rensen, M

    Bladt, M. & S rensen, M. (2005). Statistical inference for discretely observed Markov jump processes. Journal of the Royal Statistical Society: Series B 67, 395--410

  8. [8]

    Carette, P. (1995). Characterizations of embeddable 3 3 stochastic matrices with a negative eigenvalue. New York Journal of Mathematics 1, 120--129

Show all 37 references
  1. [9]

    , Falcioni, M

    Castiglione, P. , Falcioni, M. , Lesne, A. & Vulpiani, A. (2008). Chaos and coarse graining in statistical mechanics. Cambridge University Press

  2. [10]

    Connault, B. (2014). A weakly dependent Bernstein--von Mises theorem. Tech. rep., Princeton University

  3. [11]

    & Vanden-Eijnden, E

    Crommelin, D. & Vanden-Eijnden, E. (2006). Fitting timeseries by continuous-time Markov chains: A quadratic programming approach. Journal of Computational Physics 217, 782--805

  4. [12]

    & Huang, D

    Deng, K. & Huang, D. (2012). Model reduction of Markov chains via low-rank approximation. In 2012 American Control Conference (ACC). IEEE

  5. [13]

    , Pfeuffer, M

    Dos Reis, G. , Pfeuffer, M. & Smith, G. (2020). Capturing model risk and rating momentum in the estimation of probabilities of default and credit rating migrations. Quantitative Finance 20, 1069--1083

  6. [14]

    Duan, L. L. , Young, A. L. , Nishimura, A. & Dunson, D. B. (2020). Bayesian constraint relaxation. Biometrika 107, 191--204

  7. [15]

    & Sherlock, C

    Fearnhead, P. & Sherlock, C. (2006). An exact Gibbs sampler for the Markov -modulated Poisson process. Journal of the Royal Statistical Society: Series B 68, 767--784

  8. [16]

    Fill, J. A. (1991). Eigenvalue bounds on convergence to stationarity for nonreversible Markov chains, with an application to the exclusion process. The Annals of Applied Probability 1, 62--87

  9. [17]

    , Kirkpatrick, B

    Hajiaghayi, M. , Kirkpatrick, B. , Wang, L. & Bouchard-C \^o t \'e , A. (2014). Efficient continuous-time Markov chain estimation. In Proceedings of the 31st International Conference on Machine Learning. PMLR

  10. [18]

    , Meyn, S

    Huisinga, W. , Meyn, S. & Sch \"u tte, C. (2004). Phase transitions and metastability in Markovian and molecular systems. The Annals of Applied Probability 14, 419--458

  11. [19]

    Inamura, Y. et al. (2006). Estimating continuous time transition matrices from discretely observed data. Bank of Japan Working Paper 06-E07

  12. [20]

    Israel, R. B. , Rosenthal, J. S. & Wei, J. Z. (2001). Finding generators for Markov chains via empirical transition matrices, with applications to credit ratings. Mathematical finance 11, 245--265

  13. [21]

    , Hoff, P

    Jauch, M. , Hoff, P. D. & Dunson, D. B. (2020). Random orthogonal matrices and the Cayley transform . Bernoulli 26, 1560--1586

  14. [22]

    , Hoff, P

    Jauch, M. , Hoff, P. D. & Dunson, D. B. (2021). Monte Carlo simulation on the Stiefel manifold via polar expansion. Journal of Computational and Graphical Statistics 30, 622--631

  15. [23]

    Kingman, J. F. C. (1962). The imbedding problem for finite Markov chains. Zeitschrift f \"u r Wahrscheinlichkeitstheorie und verwandte Gebiete 1, 14--24

  16. [24]

    Krengel, U. (1985). Ergodic Theorems, vol. 6 of de Gruyter Studies in Mathematics. de Gruyter

  17. [25]

    Li \'e nard, J. F. & Strigul, N. S. (2016). Modelling of hardwood forest in Quebec under dynamic disturbance regimes: A time-inhomogeneous Markov chain approach. Journal of Ecology 104, 806--816

  18. [26]

    , Herring, A

    Matuk, J. , Herring, A. H. & Dunson, D. B. (2022). Bayesian modeling of nearly mutually orthogonal processes. arXiv preprint arXiv:2205.12361

  19. [27]

    , Dittmer, E

    Metzner, P. , Dittmer, E. , Jahnke, T. & Sch \"u tte, C. (2007). Generator estimation of Markov jump processes. Journal of Computational Physics 227, 353--375

  20. [28]

    Paulin, D. (2015). Concentration inequalities for Markov chains by Marton couplings and spectral methods . Electronic Journal of Probability 20, 1--32

  21. [29]

    Pfeuffer, M. (2017). ctmcd: An r package for estimating the parameters of a continuous-time markov chain from discrete-time data 9, 127--141

  22. [30]

    , Mena, R

    Riva-Palacio, A. , Mena, R. H. & Walker, S. G. (2023). On the estimation of partially observed continuous-time Markov chains. Computational Statistics 38, 1357--1389

  23. [31]

    Saunders, M. G. & Voth, G. A. (2013). Coarse-graining methods for computational biology. Annual Review of Biophysics 42, 73--93

  24. [32]

    Sherlock, C. (2021). Direct statistical inference for finite Markov jump processes via the matrix exponential. Computational Statistics 36, 2863--2887

  25. [33]

    Stewart, G. W. & Sun, J.-G. (1990). Matrix Perturbation Theory. Academic Press

  26. [34]

    Tahami Monfared, A. A. , Fu, S. , Hummel, N. , Qi, L. , Chandak, A. , Zhang, R. & Zhang, Q. (2023). Estimating transition probabilities across the Alzheimer's disease continuum using a nationally representative real-world database in the United States . Neurology and Therapy 1...

  27. [35]

    Van der Vaart, A. W. (2000). Asymptotic statistics, vol. 3. Cambridge University Press

  28. [36]

    & Wang, M

    Zhang, A. & Wang, M. (2019). Spectral state compression of Markov processes. IEEE Transactions on Information Theory 66, 3202--3231

  29. [37]

    , Wang, Z

    Zhao, T. , Wang, Z. , Cumberworth, A. , Gsponer, J. , de Freitas, N. & Bouchard-C \^o t \'e , A. (2016). Bayesian analysis of continuous time Markov chains with application to phylogenetic modelling 11, 1203--1237

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.