Exact and approximate hidden Markov chain filters based on discrete observations
-
Nicole Bäuerle
,
Igor Gilitschenski
Abstract
We consider a Hidden Markov Model (HMM) where the integrated continuous-time Markov chain can be observed at discrete time points perturbed by a Brownian motion. The aim is to derive a filter for the underlying continuous-time Markov chain. The recursion formula for the discrete-time filter is easy to derive, however involves densities which are very hard to obtain. In this paper we derive exact formulas for the necessary densities in the case the state space of the HMM consists of two elements only. This is done by relating the underlying integrated continuous-time Markov chain to the so-called asymmetric telegraph process and by using recent results on this process. In case the state space consists of more than two elements we present three different ways to approximate the densities for the filter. The first approach is based on the continuous filter problem. The second approach is to derive a PDE for the densities and solve it numerically. The third approach is a crude discrete time approximation of the Markov chain. All three approaches are compared in a numerical study.
We would like to thank our student Fehmi Mabrouk for some initial work on this topic during his Diploma Thesis which encouraged us to further pursue this topic.
References
1 N. Bäuerle and U. Rieder, Portfolio optimization with Markov-modulated stock prices and interest rates, IEEE Trans. Automat. Control 49 (2004), 442–447. 10.1109/TAC.2004.824471Search in Google Scholar
2 N. Bäuerle and U. Rieder, Markov Decision Processes with Applications to Finance, Springer, Berlin, 2011. 10.1007/978-3-642-18324-9Search in Google Scholar
3 N. Bäuerle, S. Urban and L. Veraart, The relaxed investor with partial information, SIAM J. Financial Math. 3 (2012), 304–327. 10.1137/100813646Search in Google Scholar
4 E. Bayraktar and M. Ludkovski, Sequential tracking of a hidden Markov chain using point process observations, Stochastic Process. Appl. 119 (2009), 1792–1822. 10.1016/j.spa.2008.09.003Search in Google Scholar
5 O. L. V. Costa and M. V. Araujo, A generalized multi-period mean-variance portfolio optimization with Markov switching parameters, Automatica 44 (2008), 2487–2497. 10.1016/j.automatica.2008.02.014Search in Google Scholar
6 J. Dannemann and H. Holzmann, Testing for two states in a hidden Markov model, Canad. J. Statist. 36 (2008), 505–520. 10.1002/cjs.5550360402Search in Google Scholar
7 G. B. Di Masi, Y. Kabanov and W. Runggaldier, Mean-square hedging of options on a stock with Markov volatilities, Teor. Veroyatnost. i Primenen. 39 (1994), 211–222. 10.1137/1139008Search in Google Scholar
8 R. J. Elliott, V. Krishnamurthy and J. Sass, Moment based regression algorithm for drift and volatility estimation in continuous time Markov switching models, Econometrics J. 11 (2008), 244–270. 10.1111/j.1368-423X.2008.00246.xSearch in Google Scholar
9 R. J. Elliott, L. Aggoun and J. B. Moore, Hidden Markov models, Springer, New York, 1995. Search in Google Scholar
10 R. Frey, A. Gabih and R. Wunderlich, Portfolio optimization under partial information with expert opinions, Int. J. Theor. Appl. Finance 15 (2012), 10.1142/S0219024911006486. 10.1142/S0219024911006486Search in Google Scholar
11 B. Fristedt, N. Jain and N. Krylov, Filtering and Prediction: A primer, American Mathematical Society, Providence, 2007. 10.1090/stml/038Search in Google Scholar
12 S. Frühwirth-Schnatter, Fully Bayesian analysis of switching Gaussian state space models, Ann. Inst. Statist. Math. 53 (2001), 31–49. 10.1023/A:1017908219076Search in Google Scholar
13 M. Hahn, S. Frühwirth-Schnatter and J. Sass, Markov chain Monte Carlo methods for parameter estimation in multidimensional continuous-time Markov switching models, J. Financ. Econometrics 8 (2010), 88–121. 10.1093/jjfinec/nbp026Search in Google Scholar
14 M. Hahn and J. Sass, Parameter estimation in continuous time Markov switching models: A semi-continuous Markov chain Monte Carlo approach, Bayesian Anal. 4 (2009), 63–84. 10.1214/09-BA402Search in Google Scholar
15 H. Heffes and D. Lucantoni, A Markov modulated characterization of packetized voice and data traffic and related statistical multiplexer performance, IEEE J. Select. Areas Commun. 4 (1986), 856–867. 10.1109/JSAC.1986.1146393Search in Google Scholar
16 S. M. Iacus, Statistical analysis of the inhomogeneous telegrapher's process, Statist. Probab. Lett. 55 (2001), 83–88. 10.1016/S0167-7152(01)00133-XSearch in Google Scholar
17 S. M. Iacus and N. Yoshida, Estimation for the discretely observed telegraph process, Teor. Ĭmovīr. Mat. Stat. 78 (2008), 32–42. 10.1090/S0094-9000-09-00760-1Search in Google Scholar
18 M. Kac, A stochastic model related to the telegrapher's equation, Rocky Mountain J. Math. 4 (1974), 497–509. 10.1216/RMJ-1974-4-3-497Search in Google Scholar
19 P. E. Kloeden, E. Platen and H. Schurz, Higher order approximate Markov chain filters, Stochastic Processes, Springer, New York (1993), 181–190. 10.1007/978-1-4615-7909-0_21Search in Google Scholar
20 V. G. Kulkarni, Fluid models for single buffer systems, Frontiers in Queueing, CRC, Boca Raton (1997), 321–338. Search in Google Scholar
21 O. López and N. Ratanov, On the asymmetric telegraph processes, J. Appl. Probab. 51 (2014), 569–589. 10.1017/S0021900200011438Search in Google Scholar
22 E. Otranto and G. M. Gallo, A nonparametric Bayesian approach to detect the number of regimes in Markov switching models, Econometric Rev. 21 (2002), 477–496. 10.1081/ETC-120015387Search in Google Scholar
23 E. Platen and R. Rendek, Quasi-exact approximation of hidden Markov chain filters, Commun. Stoch. Anal. 4 (2010), 129–142. 10.31390/cosa.4.1.09Search in Google Scholar
24 L. Rabehasaina and B. Sericola, A second-order Markov-modulated fluid queue with linear service rate, J. Appl. Probab. 41 (2004), 758–777. 10.1017/S0021900200020532Search in Google Scholar
25 U. Rieder and N. Bäuerle, Portfolio optimization with unobservable Markov-modulated drift process, J. Appl. Probab. 42 (2005), 362–378. 10.1239/jap/1118777176Search in Google Scholar
26 T. Rydén, An EM algorithm for estimation in Markov-modulated Poisson processes, Comput. Statist. Data Anal. 21 (1996), 431–447. 10.1016/0167-9473(95)00025-9Search in Google Scholar
27 T. Rydén, T. Teräsvirta and S. Åsbrink, Stylized facts of daily return series and the hidden Markov models, J. Appl. Econometrics 13 (1998), 217–244. 10.1002/(SICI)1099-1255(199805/06)13:3<217::AID-JAE476>3.0.CO;2-VSearch in Google Scholar
28 J. Sass and U. G. Haussmann, Optimizing the terminal wealth under partial information: The drift process as a continuous time Markov chain, Finance Stoch. 8 (2004), 553–577. 10.1007/s00780-004-0132-9Search in Google Scholar
29 J. Sass and R. Wunderlich, Optimal portfolio policies under bounded expected loss and partial information, Math. Methods Oper. Res. 72 (2010), 25–61. 10.1007/s00186-010-0300-ySearch in Google Scholar
30 R. D. Skeel and M. Berzins, A method for the spatial discretization of parabolic equations in one space variable, SIAM J. Sci. Stat. Comput. 11 (1990), 1–32. 10.1137/0911001Search in Google Scholar
31 T. Stern and A. I. Elwalid, Analysis of separable Markov-modulated rate models for information-handling systems, Adv. Appl. Probab. 1 (1991), 105–139. 10.2307/1427514Search in Google Scholar
32 Y.-C. Yao, Estimation of noisy telegraph processes: Nonlinear filtering versus nonlinear smoothing, IEEE Trans. Inform. Theory 31 (1985), 444–446. 10.1109/TIT.1985.1057032Search in Google Scholar
33 G. Yin and X. Y. Zhou, Markowitz's mean-variance portfolio selection with regime switching: From discrete-time models to their continuous-time limits, IEEE Trans. Automat. Control 49 (2004), 349–360. 10.1109/TAC.2004.824479Search in Google Scholar
34 X. Zhang, T. Siu and Q. Meng, Portfolio selection in the enlarged Markovian regime-switching market, SIAM J. Control Optim. 48 (2010), 3368–3388. 10.1137/080736351Search in Google Scholar
© 2016 by De Gruyter
