Abstract
In this paper, we establish a robustification of Elliott’s on-line EM algorithm for modelling asset prices within a hidden Markov model (HMM). In this HMM framework, parameters of the model are guided by a Markov chain in discrete time, parameters of the asset returns are therefore able to switch between different regimes. The parameters are estimated through an on-line algorithm, which utilizes incoming information from the market and leads to adaptive optimal estimates. We robustify this algorithm step by step against additive outliers appearing in the observed asset prices with the rationale to better handle possible peaks or missings in asset returns.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
- 2.
- 3.
Usually Λ θ is the logarithmic derivative of the density w.r.t. the parameter, i.e., \(\varLambda _{\theta }(x) = \partial /\partial \theta \log p_{\theta }(x)\).
References
Ang, A., Bekaert, G.: International asset allocation with regime shifts. Rev. Financ. Stud. 15, 1137–1187 (2002)
Cai, J.: A Markov model of switching-regime ARCH. J. Bus. Econ. Stat. 12, 309–316 (1994)
Elliott, R.J.: Exact adaptive filters for Markov chains observed in Gaussian noise. Automatica 30, 1399–1408 (1994)
Elliott, R.J., Hinz, J.: A method for portfolio choice. Appl. Stoch. Models Bus. Ind. 19, 1–11 (2003)
Elliott, R.J., van der Hoek, J.: An application of hidden Markov models to asset allocation problems. Financ. Stoch. 1, 229–238 (1997)
Elliott, R.J., Aggoun, L., Moore, J.B.: Hidden Markov Models: Estimation and Control. Applications of Mathematics, vol. 29. Springer, New York (1995)
Erlwein, C., Mamon, R., Davison, M.: An examination of HMM-based investment strategies for asset allocation. Appl. Stoch. Model. Bus. Ind. 27(3), 204–221 (2009)
Fernholz, L.T.: Von Mises Calculus for Statistical Functionals. Lecture Notes in Statistics, vol. 19. Spinger, New York (1983)
Fox, A.J.: Outliers in time series. J. R. Stat. Soc. Ser. B 34, 350–363 (1972)
Fraley, C., Raftery, A.E.: Model-based clustering, discriminant analysis and density estimation. J. Am. Stat. Assoc. 97, 611–631 (2002)
Fraley, C., Raftery, A.E., Murphy, T.B., Scrucca, L.: mclust version 4 for R: normal mixture modeling for model-based clustering, classification, and density estimation. Technical report No. 597, Department of Statistics, University of Washington (2012)
Gray, S.F.: Modeling the conditional distribution of interest rates as a regime-switching process. J. Financ. Econ. 42, 27–62 (1996)
Guidolin, M., Timmermann, A.: Asset allocation under multivariate regime switching. J. Econ. Dyn. Control 31, 3503–3544 (2007)
Hamilton, J.D.: A new approach to the economic analysis of nonstationary time series and the business cycle. Econom. J. Econom. Soc. 57, 357–384 (1989)
Hampel, F.R.: Contributions to the theory of robust estimation. Dissertation, University of California, Berkely (1968)
Hampel, F.R., Ronchetti, E.M., Rousseeuw, P.J. Stahel, W.A.: Robust Statistics: The Approach Based on Influence Functions. Wiley, New York (1986)
Huber, P.J.: Robust Statistics. Wiley, New York (1981)
Kohl, M.: RobLox: optimally robust influence curves and estimators for location and scale. R package version 0.8.2. http://robast.r-forge.r-project.org/ (2012)
Kohl, M., Deigner, H.P.: Preprocessing of gene expression data by optimally robust estimator. BMC Bioinform. 11, 583 (2010)
Kohl, M., Rieder, H., Ruckdeschel, P.: Infinitesimally robust estimation in general smoothly parametrized models. Stat. Methods Appl. 19, 333–354 (2010)
Knight, F.H.: Risk, Uncertainty, and Profit. Houghton Mifflin, Boston (1921)
Mamon, R., Erlwein, C., Gopaluni, B.: Adaptive signal processing of asset price dynamics with predictability analysis. Inform. Sci. 178, 203–219 (2008)
Markowitz, H.: Portfolio selection. J. Financ. 7(1), 77–91 (1952)
Maronna, R.A., Martin, R.D. Yohai, V.J.: Robust Statistics: Theory and Methods. Wiley, Chichester (2006)
Rieder, H.: Robust Asymptotic Statistics. Springer, New York (1994)
Rieder, H., Kohl, M., Ruckdeschel, P.: The cost of not knowing the radius. Stat. Methods Appl. 17(1), 13–40 (2008)
Rousseeuw, P.J., Leroy, A.M.: Robust Regression and Outlier Detection. Wiley, New York (1987)
Rousseeuw, P.J., van Driessen, K.: A fast algorithm for the minimum covariance determinant estimator. Technometrics 41, 212–223 (1999)
Ruckdeschel, P.: Ansätze zur Robustifizierung des Kalman Filters. Bayreuther Mathematische Schriften, vol. 64. Mathematisches Inst. der Univ. Bayreuth, Bayreuth (2001)
Ruckdeschel, P., Horbenko, N.: Robustness properties of estimators in generalized pareto models. Technical report ITWM No182. http://www.itwm.fraunhofer.de/fileadmin/ITWM-Media/Zentral/Pdf/Berichte_ITWM/2010/bericht_182.pdf (2010)
Sass, J., Haussmann, U.G.: Optimizing the terminal wealth under partial information: the drift process as a continuous time Markov chain. Financ. Stoch. 8, 553–577 (2004)
Zakai, M.: On the optimal filtering of diffusion processes. Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete 11, 230–243 (1969)
Acknowledgements
We thank an anonymous referee and the editor for valuable comments. Financial support for C. Erlwein-Sayer from Deutsche Forschungsgemeinschaft (DFG) within the project “Regimeswitching in zeitstetigen Finanzmarktmodellen: Statistik und problemspezifische Modellwahl” (RU-893/4-1) is gratefully acknowledged.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix
Appendix
1.1.1 Definition: Weighted Medians and MADs
For weights w j ≥ 0 and observations y j , the weighted median m = m(y, w) is defined as
With \(y^{\prime}_{j} = \vert y_{j} - m\vert \), the scaled weighted MAD s = s(y, w) is defined as
where c is a consistency factor. It warrants consistent estimation of σ in case of Gaussian observations, i.e., \(c = \mbox{ E}\,\mathrm{argmin}_{t}\sum _{j}w_{j}\vert \vert \tilde{y}_{j}\vert - t\vert \) for \(\tilde{y}_{j}\stackrel{\mathrm{i.i.d.}}{\sim }\mathcal{N}(0,1)\). c can be obtained empirically for a sufficiently large sample size M, e.g., M = 10, 000, setting \(c = \frac{1} {M}\sum _{k=1}^{M}c_{ k}\), \(c_{k} = \mathrm{argmin}_{t}\sum _{j}w_{j}\vert \vert y^{\prime\prime}_{j,k}\vert - t\vert \), \(y^{\prime\prime}_{j,k}\stackrel{\mathrm{i.i.d.}}{\sim }\mathcal{N}(0,1)\).
As to the (finite sample) breakdown point FSBP of the weighted median (and at the same time for the scaled weighted MAD), we define \(w_{j}^{0} = w_{i,j}/\sum _{j^{\prime}}w_{j^{\prime}}\), and for each i define the ordered weights w (j) 0 such that w (1) 0 ≥ w (2) 0 ≥ … ≥ w (k) 0. Then the FSBP in both cases is \({k}^{-1}\min \{j_{0} = 1,\ldots,k\,\mid \,\sum _{j=1}^{j_{0}}w_{(j)}^{0} \geq k/2\}\) which (for equivariant estimators) can be shown to be the largest possible value. So using weighted medians and MADs, we achieve a decent degree of robustness against outliers. E.g., assume we have 10 observations with weights 5 × 0. 05; 3 × 0. 1; 0. 2; 0. 25. Then we need at least three outliers (placed at weights 0. 1, 0. 2, 0. 25, respectively) to produce a breakdown.
1.1.2 Proof of Theorem 1.2
Proof.
Let us solve \(\max _{\partial \mathcal{U}}\min _{f}[\ldots ]\) first, which amounts to \(\min _{\partial \mathcal{U}}\mbox{ E}_{\mathrm{re}}[\vert \mbox{ E}_{\mathrm{re}}[{Y }^{\mathrm{id}}\vert {Y }^{\mathrm{re}}]{\vert }^{2}]\). For fixed element \({P}^{{Y }^{\mathrm{di}} }\) assume a dominating σ-finite measure μ, i.e., \(\mu \gg {P}^{{Y }^{\mathrm{di}} }\), \(\mu \gg {P}^{{Y }^{\mathrm{id}} }\); this gives us a μ-density q(y) of \({P}^{{Y }^{\mathrm{di}} }\). Determining the joint (real) law \({P}^{{Y }^{\mathrm{id}},{Y }^{\mathrm{re}} }(d\tilde{y},dy)\) as
we deduce that μ(dy)-a.e.
Hence we have to minimize
in \(M_{0} =\{ q \in L_{1}(\mu )\,\vert \;q \geq 0,\;\int q\,d\mu = 1\}\). To this end, we note that F is convex on the non-void, convex cone M = { q ∈ L 1(μ) | q ≥ 0} so, for some \(\tilde{\rho }\geq 0\), we may consider the Lagrangian
for some positive Lagrange multiplier \(\tilde{\rho }\). Pointwise minimization in y of \(L_{\tilde{\rho }}(q)\) gives
for some constant \(s = s(\tilde{\rho }) = {(\,\vert \mbox{ E}{Y }^{\mathrm{}id}{\vert }^{2} +\tilde{\rho } /r)}^{1/2}\), Pointwise in y, \(\hat{q}_{s}\) is antitone and continuous in s ≥ 0 and \(\lim _{s\rightarrow 0[\infty ]}q_{s}(y) = \infty [0]\), hence by monotone convergence,
too, is antitone and continuous and \(\lim _{s\rightarrow 0[\infty ]}H(s) = \infty [0]\). So by continuity, there is some ρ ∈ (0, ∞) with H(ρ) = 1. On M 0, ∫ q d μ = 1, but \(\hat{q}_{\rho } = q_{s=\rho } \in M_{0}\) and is optimal on M ⊃ M 0 hence it also minimizes F on M 0. In particular, we get representation (1.26) and note that, independently from the choice of μ, the least favorable \(P_{0}^{{Y }^{\mathrm{di}} }\) is dominated according to \(P_{0}^{{Y }^{\mathrm{di}} } \ll {P}^{{Y }^{\mathrm{id}} }\), i.e.; non-dominated \({P}^{{Y }^{\mathrm{di}} }\) are even easier to deal with.
As next step we show that
To this end we first verify (1.25) determining f 0(y) as \(f_{0}(y) = \mbox{ E}_{\mathrm{re};\hat{P}}[X\vert {Y }^{\mathrm{re}} = y]\). Writing a sub/superscript “re; P” for evaluation under the situation generated by \(P = {P}^{{Y }^{\mathrm{di}} }\) and \(\hat{P}\) for \(P_{0}^{{Y }^{\mathrm{di}} }\), we obtain the risk for general P as
This is maximal for any P that is concentrated on the set \(\big\{\,\vert D({Y }^{\mathrm{di;},q})\vert >\rho \,\big\}\), which is true for \(\hat{P}\). Hence (1.49) follows, as for any contaminating P
Finally, we pass over from \(\partial \mathcal{U}\) to \(\mathcal{U}\): Let f r , \(\hat{P}_{r}\) denote the components of the saddle-point for \(\partial \mathcal{U}(r)\), as well as ρ(r) the corresponding Lagrange multiplier and w r the corresponding weight, i.e., \(w_{r} = w_{r}(y) =\min (1,\rho (r)\,/\,\vert D(y)\vert )\). Let R(f, P, r) be the MSE of procedure f at the SO model \(\partial \mathcal{U}(r)\) with contaminating \({P}^{{Y }^{\mathrm{di}} } = P\). As can be seen from (1.26), ρ(r) is antitone in r; in particular, as \(\hat{P}_{r}\) is concentrated on { | D(Y ) | ≥ ρ(r)} which for r ≤ s is a subset of { | D(Y ) | ≥ ρ(s)}, we obtain
Note that R(f s , P, 0) = R(f s , Q, 0) for all P, Q – hence passage to \(\tilde{R}(f_{s},P,r) = R(f_{s},P,r) - R(f_{s},P,0)\) is helpful – and that
Abbreviate \(\bar{w}_{s}({Y }^{\mathrm{id}}) = 1 - {(1 - w_{ s}({Y }^{\mathrm{id}}))}^{2} \geq 0\) to see that
Hence the saddle-point extends to \(\mathcal{U}(r)\); in particular the maximal risk is never attained in the interior \(\mathcal{U}(r)\setminus \partial \mathcal{U}(r)\). (1.28) follows by plugging in the results. □
Rights and permissions
Copyright information
© 2014 Springer Science+Business Media New York
About this chapter
Cite this chapter
Erlwein-Sayer, C., Ruckdeschel, P. (2014). Robustification of an On-line EM Algorithm for Modelling Asset Prices Within an HMM. In: Mamon, R., Elliott, R. (eds) Hidden Markov Models in Finance. International Series in Operations Research & Management Science, vol 209. Springer, Boston, MA. https://doi.org/10.1007/978-1-4899-7442-6_1
Download citation
DOI: https://doi.org/10.1007/978-1-4899-7442-6_1
Published:
Publisher Name: Springer, Boston, MA
Print ISBN: 978-1-4899-7441-9
Online ISBN: 978-1-4899-7442-6
eBook Packages: Business and EconomicsBusiness and Management (R0)