Page 217 - Special Topic Session (STS) - Volume 1
P. 217
STS425 Nur I. et al.
4. Discussion and Conclusion
The existence of EM as a numerical estimation method in the estimation
of the parameters of the MSwM (.) - AR (.) model can be shown to be able to
help us in estimating the first step visit of each regime. This first visit can be
calculated as a run length for the raised regime. Therefore, ARL can finally be
obtained as a result of a side process of EM during the parameter estimation
of MSw () - AR (). The results of the run length and ARL recording process can
be used as a material for monitoring stock price movements for day-to-day
transactions. At the end of the series which AALI had been removed from LQ45
during the last run length, AALI backed to the first regime as the period before
it listed in LQ45 at 2003. Investments in stocks such as AALI will tend to be
easier to monitor the movements of the share than on stocks such as SSMS,
which changes the regime quickly.
The estimation method with EM is always constrained by the number of
regimes that must be predetermined first. This is a new challenge if there is a
sudden process that will change the number of regimes when this method is
applied to monitoring in real observation.
References
1. Britama (2019) Daftar Saham dalam Indeks LQ-45 Mulai dari Agustus
2003 hingga Sekarang, http://britama.com/index.php/2016/03/daftar-
saham-dalam-indeks-lq-45-mulai-dari-agustus-2003-hingga-sekarang/,
downloaded on April 10, 2019
2. Chuffart, T. (2015) Selection Criteria in Regime Switching Conditional
Volatility Models, Econometrics, 2015, 3, 289-316.
DOI:10.3390/econometrics3020289.
3. Dempster, A.P., Laird, N.M. & Rubin, D.B. (1977) Maximum likelihood
from incomplete data via the EM algorithm (with discussion). Journal of
the Royal Statistical Society Series B, 39, 1-38. DOI: 10.1.1.133.4884.
4. Frühwirth-Schnatter, S. (2006) Finite Mixture and Markov Switching
Models, Springer, New York.
5. Hamilton, J.D. (1996) Specification Testing in Markov-switching Time
Series Models. Journal of Econometrics, 70: 127-157.
6. Huang, X., Xu, N., & Bisgaard, S. (2013) A Class of Markov Chain Models
for Average Run Length Computations for Autocorrelated Processes,
Communications in Statistics - Simulation and Computation, 42:7, 1495-
1513, DOI: 10.1080/03610918.2012.667474
7. Kim, C.J & Nelson C.R. (1999) State Space Models with Regime
Switching, Classical and Gibbs Sampling Approaches with Applications.
Cambridge, MA: MIT Press.
206 | I S I W S C 2 0 1 9