Computational issues in parameter estimation for hidden Markov models with template model builder

References

• Juang BH, Rabiner LR. Hidden Markov models for speech recognition. Technometrics. 1991 Aug;33(3):251–272. doi: 10.1080/00401706.1991.10484833
   Web of Science ®Google Scholar
• Baum LE, Petrie T. Statistical inference for probabilistic functions of finite state Markov chains. Ann Math Stat. 1966 Dec;37(6):1554–1563. doi: 10.1214/aoms/1177699147
   Google Scholar
• Rabiner L, Juang B. An introduction to hidden Markov models. IEEE ASSP Mag. 1986 Jan;3(1):4–16. doi: 10.1109/MASSP.1986.1165342
   Google Scholar
• Rabiner L. A tutorial on hidden Markov models and selected applications in speech recognition. Proc IEEE. 1989 Feb;77(2):257–286. doi: 10.1109/5.18626
   Web of Science ®Google Scholar
• Fredkin DR, Rice JA. Bayesian restoration of single-channel patch clamp recordings. Biometrics. 1992;48(2):427–448. doi: 10.2307/2532301
   PubMed Web of Science ®Google Scholar
• Schadt EE, Sinsheimer JS, Lange K. Computational advances in maximum likelihood methods for molecular phylogeny. Genome Res. 1998 Jan;8(3):222–233. doi: 10.1101/gr.8.3.222
   PubMed Web of Science ®Google Scholar
• Durbin R. Biological sequence analysis: probabilistic models of proteins and nucleic acids. Cambridge: Cambridge University Press; 1998.
   Google Scholar
• Eddy SR. Profile hidden Markov models. Bioinformatics. 1998 Jan;14(9):755–763. doi: 10.1093/bioinformatics/14.9.755
   PubMed Web of Science ®Google Scholar
• Hamilton JD. A new approach to the economic analysis of nonstationary time series and the business cycle. Econometrica. 1989;57(2):357–384. doi: 10.2307/1912559
   Web of Science ®Google Scholar
• McClintock BT, Langrock R, Gimenez O, et al. Uncovering ecological state dynamics with hidden Markov models. Ecol Lett. 2020;23(12):1878–1903. doi: 10.1111/ele.v23.12
   PubMed Web of Science ®Google Scholar
• Lystig TC, Hughes JP. Exact computation of the observed information matrix for hidden Markov models. J Comput Graph Stat. 2002 Sep;11(3):678–689. doi: 10.1198/106186002402
   Web of Science ®Google Scholar
• Ailliot P, Allard D, Monbet V, et al. Stochastic weather generators: an overview of weather type models. J Soc Fr Stat. 2015;156(1):101–113.
   Google Scholar
• Mor B, Garhwal S, Kumar A. A systematic review of hidden Markov models and their applications. Arch Comput Methods Eng. 2021 May;28(3):1429–1448. doi: 10.1007/s11831-020-09422-4
   Web of Science ®Google Scholar
• Baum LE, Petrie T, Soules G, et al. A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains. Ann Math Stat. 1970;41(1):164–171. doi: 10.1214/aoms/1177697196
   Google Scholar
• Dempster AP, Laird NM, Rubin DB. Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc Ser B (Methodol). 1977;39(1):1–22. doi: 10.1111/j.2517-6161.1977.tb01600.x
   Google Scholar
• Liporace L. Maximum likelihood estimation for multivariate observations of Markov sources. IEEE Trans Inform Theory. 1982 Sep;28(5):729–734. doi: 10.1109/TIT.1982.1056544
   Web of Science ®Google Scholar
• Wu CFJ. On the convergence properties of the EM algorithm. The Ann Stat. 1983;11(1):95–103. doi: 10.1214/aos/1176346060
   Web of Science ®Google Scholar
• Turner R. Direct maximization of the likelihood of a hidden Markov model. Comput Stat Data Anal. 2008 May;52(9):4147–4160. doi: 10.1016/j.csda.2008.01.029
   Web of Science ®Google Scholar
• MacDonald IL, Zucchini W. Hidden Markov and other models for discrete-valued time series. London: Chapman & Hall; 1997. (Monographs on statistics and applied probability; vol. 70).
   Google Scholar
• Cappé O, Moulines E, Ryden T. Inference in hidden Markov models. Berlin: Springer Science & Business Media; 2005.
   Google Scholar
• Lange K, Weeks DE. Efficient computation of lod scores: genotype elimination, genotype redefinition, and hybrid maximum likelihood algorithms. Ann Hum Genet. 1989;53(1):67–83. doi: 10.1111/ahg.1989.53.issue-1
   PubMed Web of Science ®Google Scholar
• Zucchini W, MacDonald I, Langrock R. Hidden markov models for time series: an introduction using R. 2nd ed. CRC Press; 2016. (Chapman & Hall/CRC Monographs on Statistics & Applied Probability).
   Google Scholar
• Altman RM, Petkau AJ. Application of hidden Markov models to multiple sclerosis lesion count data. Stat Med. 2005;24(15):2335–2344. doi: 10.1002/(ISSN)1097-0258
   PubMed Web of Science ®Google Scholar
• MacDonald IL. Numerical maximisation of likelihood: a neglected alternative to EM? Int Stat Rev. 2014 Aug;82(2):296–308. doi: 10.1111/insr.v82.2
   Web of Science ®Google Scholar
• Kristensen K, Nielsen A, Berg C, et al. TMB: automatic differentiation and Laplace approximation. J Stat Softw Articles. 2016;70(5):1–21. doi: 10.18637/jss.v070.i05
   Web of Science ®Google Scholar
• Bacri T, Berentsen GD, Bulla J, et al. A gentle tutorial on accelerated parameter and confidence interval estimation for hidden Markov models using template model builder. Biom J. 2022 Oct;64(7):1260–1288. doi: 10.1002/bimj.v64.7
   PubMed Web of Science ®Google Scholar
• Gay DM. Usage summary for selected optimization routines. Computing Science Technical Report. 1990;153:1–21.
   Google Scholar
• Nash JC, Varadhan R. Optimr: a replacement and extension of the optim function; 2019. R package version 2019-12.16; Available from: https://CRAN.R-project.org/package=optimr.
   Google Scholar
• Bulla J, Berzel A. Computational issues in parameter estimation for stationary hidden Markov models. Comput Stat. 2008 Jan;23(1):1–18. doi: 10.1007/s00180-007-0063-y
   Web of Science ®Google Scholar
• Dennis Jr JE, Moré JJ. Quasi-Newton methods, motivation and theory. SIAM Rev. 1977 Jan;19(1):46–89. doi: 10.1137/1019005
   Web of Science ®Google Scholar
• Schnabel RB, Koonatz JE, Weiss BE. A modular system of algorithms for unconstrained minimization. ACM Trans Math Softw. 1985 Dec;11(4):419–440. doi: 10.1145/6187.6192
   Web of Science ®Google Scholar
• Nelder JA, Mead R. A simplex method for function minimization. Comput J. 1965 Jan;7(4):308–313. doi: 10.1093/comjnl/7.4.308
   Web of Science ®Google Scholar
• Hassan MR, Nath B, Kirley M. A fusion model of HMM, ANN and GA for stock market forecasting. Expert Syst Appl. 2007 Jul;33(1):171–180. doi: 10.1016/j.eswa.2006.04.007
   Web of Science ®Google Scholar
• Stolcke A, Omohundro S. Hidden Markov model induction by Bayesian model merging. Adv Neural Inf Process Syst. 1993;11–11.
   Google Scholar
• Oudelha M, Ainon RN. HMM parameters estimation using hybrid Baum-Welch genetic algorithm. In: 2010 International Symposium on Information Technology; Vol. 2; Jun.; 2010. p. 542–545. Kuala Lumpur, Malaysia
   Google Scholar
• Redner RA, Walker HF. Mixture densities, maximum likelihood and the EM algorithm. SIAM Rev. 1984 Apr;26(2):195–239. doi: 10.1137/1026034
   Web of Science ®Google Scholar
• Kato J, Watanabe T, Joga S, et al. An HMM-based segmentation method for traffic monitoring movies. IEEE Trans Pattern Anal Mach Intell. 2002 Sep;24(9):1291–1296. doi: 10.1109/TPAMI.2002.1033221
   Web of Science ®Google Scholar
• Maruotti A, Punzo A. Initialization of hidden Markov and semi-Markov models: A critical evaluation of several strategies. Int Stat Rev. 2021 Dec;89(3):447–480. doi: 10.1111/insr.v89.3
   Web of Science ®Google Scholar
• Michelot T. hmmTMB: Hidden Markov models with flexible covariate effects in R; 2023.
   Google Scholar
• Feller W. An introduction to probability theory and its applications. New York: Wiley; 1968.
   Google Scholar
• Pryss R, Reichert M, Langguth B, et al. Mobile crowd sensing services for tinnitus assessment, therapy, and research. In: 2015 IEEE International Conference on Mobile Services; Jun. IEEE; 2015. p. 352–359.
   Google Scholar
• Pryss R, Reichert M, Herrmann J, et al. Mobile crowd sensing in clinical and psychological trials–a case study. In: 2015 IEEE 28th International Symposium on Computer-Based Medical Systems; Jun. IEEE; 2015. p. 23–24.
   Google Scholar
• Probst T, Pryss R, Langguth B, et al. Emotion dynamics and tinnitus: Daily life data from the ‘TrackYourTinnitus’ application. Sci Rep. 2016 Aug;6(1):31166. doi: 10.1038/srep31166
   PubMed Web of Science ®Google Scholar
• Probst T, Pryss RC, Langguth B, et al. Does tinnitus depend on time-of-day? An ecological momentary assessment study with the ‘TrackYourTinnitus’ application. Front Aging Neurosci. 2017 Aug;9:253. doi: 10.3389/fnagi.2017.00253
   PubMed Web of Science ®Google Scholar
• Rydén T, Teräsvirta T, Åsbrink S. Stylized facts of daily return series and the hidden Markov model. J Appl Econom. 1998;13(3):217–244. doi: 10.1002/(ISSN)1099-1255
   Web of Science ®Google Scholar
• Schwert GW. Why does stock market volatility change over time. J Finance. 1989;44(5):1115–1153. doi: 10.1111/j.1540-6261.1989.tb02647.x
   Web of Science ®Google Scholar
• Maheu John M, McCurdy Thomas H. Identifying bull and bear markets in stock returns. J Bus Econ Stat. 2000 Jan;18(1):100–112. doi: 10.1080/07350015.2000.10524851.
   Web of Science ®Google Scholar
• Guidolin M, Timmermann A. Economic implications of bull and bear regimes in UK stock and bond returns. Econom J. 2005 Jan;115(500):111–143. doi: 10.1111/j.1468-0297.2004.00962.x
   Web of Science ®Google Scholar
• R Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2021.
   Google Scholar
• Schwendinger F, Borchers HW. CRAN task view: optimization and mathematical programming. 2022. [Available from: https://cran.r-project.org/view=optimization].
   Google Scholar
• Broyden CG. The convergence of a class of double-rank minimization algorithms 1. General considerations. IMA J Appl Math. 1970 Mar;6(1):76–90. doi: 10.1093/imamat/6.1.76
   Google Scholar
• Fletcher R. A new approach to variable metric algorithms. Comput J. 1970 Jan;13(3):317–322. doi: 10.1093/comjnl/13.3.317
   Web of Science ®Google Scholar
• Goldfarb D. A family of variable-metric methods derived by variational means. Math Comput. 1970;24(109):23–26. doi: 10.1090/mcom/1970-24-109
   Web of Science ®Google Scholar
• Shanno DF. Conditioning of quasi-Newton methods for function minimization. Math Comput. 1970;24(111):647–656. doi: 10.1090/mcom/1970-24-111
   Web of Science ®Google Scholar
• Byrd RH, Lu P, Nocedal J, et al. A limited memory algorithm for bound constrained optimization. SIAM J Sci Comput. 1995 Sep;16(5):1190–1208. doi: 10.1137/0916069
   Web of Science ®Google Scholar
• Fletcher R. Practical methods of optimization. Hoboken, New Jersey, U.S.A: John Wiley & Sons; 2013.
   Google Scholar
• Dennis JE, Schnabel RB. Numerical methods for unconstrained optimization and nonlinear equations. Society for Industrial and Applied Mathematics; 1996. (Classics in Applied Mathematics).
   Google Scholar
• Hooke R, Jeeves TA. ‘Direct search’ solution of numerical and statistical problems. J ACM. 1961 Apr;8(2):212–229. doi: 10.1145/321062.321069
   Google Scholar
• Philipps V, Proust-Lima C, Prague M, et al. Marqlevalg: a parallelized general-purpose optimization based on marquardt-levenberg algorithm; 2023. R package version 2.0.8; Available from: https://CRAN.R-project.org/package=marqLevAlg.
   Google Scholar
• Levenberg K. A method for the solution of certain non-linear problems in least squares. Quart Appl Math. 1944;2(2):164–168. doi: 10.1090/qam/1944-02-02
   Google Scholar
• Marquardt DW. An algorithm for least-squares estimation of nonlinear parameters. J Soc Indust Appl Math. 1963 Jun;11(2):431–441. doi: 10.1137/0111030
   Google Scholar
• Varadhan R, Gilbert P. Bb: solving and optimizing large-scale nonlinear systems; 2019. R package version 2019.10-1; Available from: http://www.jhsph.edu/agingandhealth/People/Faculty_personal_pages/Varadhan.html.
   Google Scholar
• Varadhan R, Gilbert P. BB: an R package for solving a large system of nonlinear equations and for optimizing a high-dimensional nonlinear objective function. J Stat Softw. 2010;32:1–26. doi: 10.18637/jss.v032.i04
   Web of Science ®Google Scholar
• Bates D, Mullen KM, Nash JC, et al. minqa: derivative-free optimization algorithms by quadratic approximation; 2022. R package version 1.2.5; Available from: http://optimizer.r-forge.r-project.org.
   Google Scholar
• Powell MJD. The NEWUOA software for unconstrained optimization without derivatives. In: Di Pillo G, Roma M, editors. Large-scale nonlinear optimization. Boston (MA): Springer US; 2006. p. 255–297. (Nonconvex Optimization and Its Applications).
   Google Scholar
• Mersmann O. microbenchmark: accurate timing functions; 2023. R package version 1.4.10; Available from: https://github.com/joshuaulrich/microbenchmark/.
   Google Scholar
• Böhning D. Computer-assisted analysis of mixtures and applications: meta-analysis, disease mapping and others. Vol. 81. Boca Raton, Florida: CRC press; 1999.
   Google Scholar
• Bhatia R, Davis C. A better bound on the variance. Am Math Mon. 2000 Apr;107(4):353–357. doi: 10.1080/00029890.2000.12005203
   Web of Science ®Google Scholar
• Nagy J. Über algebraische Gleichungen mit lauter reellen Wurzeln. Jahresbericht Der Deutschen Mathematiker-Vereinigung. 1918;27:37–43.
   Google Scholar
• Biernacki C, Celeux G, Govaert G. Assessing a mixture model for clustering with the integrated completed likelihood. IEEE Trans Pattern Anal Mach Intell. 2000 Jul;22(7):719–725. doi: 10.1109/34.865189
   Web of Science ®Google Scholar
• Visser I, Raijmakers MEJ, Molenaar PCM. Confidence intervals for hidden Markov model parameters. British J Math Stat Psychol. 2000;53(2):317–327. doi: 10.1348/000711000159240
   PubMed Web of Science ®Google Scholar
• Pawitan Y. In all likelihood: statistical modelling and inference using likelihood. Oxford: Oxford University Press; 2001.
   Google Scholar
• Zheng N, Cadigan N. Frequentist delta-variance approximations with mixed-effects models and TMB. Comput Stat Data Anal. 2021 Aug;160:107227. doi: 10.1016/j.csda.2021.107227
   Web of Science ®Google Scholar
• Efron B, Tibshirani RJ. An introduction to the bootstrap. New York (NY); London: Chapman & Hall; 1993.
   Google Scholar
• Härdle W, Horowitz J, Kreiss JP. Bootstrap methods for time series. Int Stat Revi. 2003;71(2):435–459. doi: 10.1111/insr.2003.71.issue-2
   Web of Science ®Google Scholar