Abstract
The multivariate continuous-time ARMA model is a tool to capture the relationship between multivariate time series. In this chapter, a particular computational implementation of a stationary normal multivariate CARMA model is illustrated. A review of a parametric setup is shown. Data are assumed to be observed at irregular non-synchronous discrete time points. The computational approach for calculating the likelihood is based on a state-space form and the Kalman filter. Interpretation of the CARMA models is discussed. The computational algorithms have been implemented in R packages. Examples of a simulated and real data are shown.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Belcher, J., Hampton, J., & Tunnicliffe Wilson, G. (1994). Parameterization of continuous autoregressive models for irregularly sampled time series data. Journal of the Royal Statistical Association, Series B, 56(1), 141–155.
Björk, T. (1998). Arbitrage theory in continuous time. Oxford: Oxford University Press. https://doi.org/10.1093/0198775180.001.0001
Box, G. E. P., & Jenkins, G. M. (1976). Time series analysis: Forecasting and control. San Fransisco: Holden Day.
Brockwell, P. J. (2014). Recent results in the theory and applications of carma processes. Annals of the Institute of Statistical Mathematics, 66(4), 647–685. https://doi.org/10.1007/s10463-014-0468-7
Brockwell, P. J., & Davis, R. A. (1991). Time series: Theory and methods. New York: Springer. https://doi.org/10.1007/978-1-4419-0320-4
Brockwell, P. J., & Marquardt, T. (2005). Levy-driven and fractionally integrated ARMA processes with continuous time parameter. Statistica Sinica, 15, 477–494.
Fasen, V., & Fuchs, F. (2013). Spectral estimates for high-frequency sampled continuous-time autoregressive moving average processes. Journal of Time Series Analysis, 34(5), 532–551. https://doi.org/10.1111/jtsa.12029
Garcia, I., Klüppelberg, C., & Müller, G. (2011). Estimation of stable CARMA models with an application to electricity spot prices. Statistical Modelling, 11(5), 447–470. https://doi.org/10.1177/1471082X1001100504
Greene, W. H. (2012). Econometric analysis (7th ed.). Upper Saddle River, NJ: Pearson.
Harvey, A. C. (1989). Forecasting, structural time series models and the Kalman filter. Cambridge: Cambridge University Press. https://doi.org/10.1017/cbo9781107049994
Iacus, S. M., & Mercuri, L. (2015). Implementation of Lévy Carma model in Yuima package. Computational Statistics, 30(4), 1111–1141. https://doi.org/10.1007/s00180-015-0569-7
Jouzel, J., et al. (2007). EPICA Dome C ice core 800kyr deuterium data and temperature estimates. IGBP PAGES/World Data Center for Paleoclimatology Data Contribution Series # 2007-091. NOAA/NCDC Paleoclimatology Program, Boulder CO, USA.
Kawai, R. (2015). Sample path generation of Lévy-driven continuous-time autoregressive moving average processes. Methodology and Computing in Applied Probability, 19(1), 175–211. https://doi.org/10.1007/s11009-015-9472-5
Kreyszig, E. (1999). Advanced engineering mathematics (8th ed.). London: Wiley (Linear fractional transformation, p. 692).
Luthi, D., Floch, M. L., Bereiter, B., Blunier, T., Barnola, J.-M., Siegenthaler, U., et al. (2008). High-resolution carbon dioxide concentration record 650,000-800,000 years before present. Nature, 453, 379–382. https://doi.org/10.1038/nature06949
Lütkepohl, H. (1991). Introduction to multiple time series analysis. Berlin: Springer. https://doi.org/10.1007/978-3-662-02691-5
Øksendal, B. (1998). Stochastic differential equations: An introduction with applications (5th ed.). Berlin: Springer. https://doi.org/10.1007/978-3-662-03620-4
Oud, J. H. L., & Voelkle, M. C. (2014). Do missing values exist? incomplete data handling in cross-national longitudinal studies by means of continuous time modeling. Quality & Quantity, 48(6), 3271–3288. https://doi.org/10.1007/s11135-013-9955-9
Schlemm, E., & Stelzer, R. (2012). Multivariate CARMA processes, continuous-time state space models and complete regularity of the innovations of the sampled processes. Bernoulli, 18(1), 46–64. https://doi.org/10.3150/10-BEJ329
Singer, H. (1992). Continuous-time dynamical systems with sampled data, errors of measurment and unobserved components. Journal of Time Series, 14(5), 527–544. https://doi.org/10.1111/j.1467-9892.1993.tb00162.x
Singer, H. (1995). Analytical score function for irregularly sampled continuous time stochastic processes with control variables and missing values. Econometric Theory, 11(4), 721–735. https://doi.org/10.1017/S0266466600009701
Tómasson, H. (2015). Some computational aspects of Gaussian CARMA modelling. Statistics and Computing, 25(2), 375–387. https://doi.org/10.1007/s11222-013-9438-9
Tsai, H., & Chan, K. (2000). A note on the covariance structure of a continuous-time ARMA process. Statistica Sinica, 10, 989–998.
Zadrozny, P. (1988). Gaussian likelihood of continuous-time ARMAX models when data are stocks and flows at different frequencies. Econometric Theory, 4(1), 108–124. https://doi.org/10.1017/S0266466600011890
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendices
Appendix 1: Implementation of Kalman Filter
The likelihood function is calculated by use of the Kalman-filter recursions. The dynamics are defined by Eqs. (15.15) and (15.16). The sample selection matrix C contains 1 at coordinate ii when coordinate i is observed at time point t i and is zero otherwise. Given information at time t i−1,the optimal estimate of the state vector is X(t i−1|t i−1) and the corresponding variance matrix P X(t i−1|t i−1). The prediction step is:
The corresponding updating equations:
Here F ∗(t i|t i−1) denotes a generalised inverse of F(t i|t i−1). The normal likelihood is easily calculated in the non-synchronous case, because at each time point, observed data are essentially univariate. The observed value is compared to its predicted value. The corresponding variance is the non-zero element of element of F(t i|t i−1)).
Appendix 2: A Brief Description of the R Packages
The R packages are based on objects. In the ctarmaRcpp package, a ctarma object contains the observed series, y=y(t 1), …, y(t n); the time points of observations, tt=t 1, …, t n; the AR parameters, a=(α 1, …, α p); the MA parameters b=(β 1, …, β q); and the standard deviation of the innovations, sigma=σ.
names(ctarma1) [1] "y" "tt" "a" "b" "sigma" >ctarma1$y[1:5] [1] -1.330619311 0.215559470 0.218091305 -0.004512067 -0.124261787 >ctarma1$tt[1:5] [1] 0.08360339 0.51537932 0.65898476 0.77687963 0.86278019 >> ctarma1$a [1] 2 40 > ctarma1$b [1] 1.00 0.15 >ctarma1$sigma [1] 8
The package ctarmaRcpp contains functions on ctarma objects for, e.g., calculating and maximising the log-likelihood. Spectral densities can also be calculated and plotted.
In the multivariate mctarmaRcpp, a second compatible (same number of AR and MA coefficients) ctarma object ctarma2 can be merged with ctarma1 into a bivariate mctarma object.
> cc=cbind(as.list(ctarma0),as.list(ctarma2)) > mctobj=mctobjnonsync(cc) > mctobj$A [,1] [,2] [,3] [,4] [1,] -2 0 1 0 [2,] 0 -2 0 1 [3,] -40 0 0 0 [4,] 0 -40 0 0 > mctobj$R [,1] [,2] [1,] 0.15 0.00 [2,] 0.00 0.15 [3,] 1.00 0.00 [4,] 0.00 1.00 > mctobj$sigma [,1] [,2] [1,] 64 0 [2,] 0 64
That is, the multivariate object mctobj now contains the matrix of AR parameters A, the matrix of MA parameters R and the variance matrix of the innovations Σ. The object also contains a vector measurements y, an ordered vector of time points of measurements tt and a vector z denoting which coordinate of the y vector was observed.
cbind(mctobj$y,mctobj$tt,mctobj$z)[1:5,] [,1] [,2] [,3] [1,] -0.6597594 0.03740642 2 [2,] -1.3306193 0.08360339 1 [3,] 0.9946577 0.27623398 2 [4,] 1.1272836 0.31624703 2 [5,] 1.2388762 0.32568241 2
The observations are non-synchronous, i.e. only one coordinate of the y vector is observed at each time point. Observations 1 and 3–5 are of coordinate 2; the second is of coordinate 1.
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG, part of Springer Nature
About this chapter
Cite this chapter
Tómasson, H. (2018). Implementation of Multivariate Continuous-Time ARMA Models. In: van Montfort, K., Oud, J.H.L., Voelkle, M.C. (eds) Continuous Time Modeling in the Behavioral and Related Sciences. Springer, Cham. https://doi.org/10.1007/978-3-319-77219-6_15
Download citation
DOI: https://doi.org/10.1007/978-3-319-77219-6_15
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-77218-9
Online ISBN: 978-3-319-77219-6
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)