Guaranteed estimation of parameters of threshold autoregressive process with conditional heteroskedasticity
A nonlinear autoregressive model TAR/ARCH(1,1) was considered. In the model the value of the autoregressive parameter and the noise variation depend on the value of the process in the previous moment. Sufficient conditions of the ergodicity of the process were obtained. Estimators of the autoregressive parameters if all the process parameters are unknown were constructed. The estimators are unbiased and their variations are bounded from above by values depending of the estimation procedure parameter. Asymptotic properties of the estimators were investigated.
Download file
Counter downloads: 431
Keywords
TAR/ARCH, метод наименьших квадратов, среднеквад-ратическое отклонение, гарантированное оценивание, TAR/ARCH, least squares method, mean square error, guaranteed estimationAuthors
| Name | Organization | |
| Burkatovskaya Yulia B. | Tomsk Polytechnic University; Tomsk State University | tracey@tpu.ru |
| Vorobeychikov Sergey E. | Tomsk State University | sev@mail.tsu.ru |
References
Joseph D. Petrucelli, Samuel W. Woolford. A threshold AR(1) model // J. Appl. Prob. 1984. V. 21. P. 270-286.
Rong Chen, Ruey S. Tsay. On the ergodicity of TAR(1) model // The Annals of Applied Probability. 1991. V. 1. No. 4. P. 613-634.
Sangyeol Lee, T.N. Shtram. Sequential point estimation of parameters in a threshold AR (1) model // Stochastic Processes and Their Applications. 1999. V. 84. P. 343-355.
Lee O., Shin D.W. A note on the geometric ergodicity of a multiple threshold AR(1) processes on the boundary region with application to integrated m-m processes.// Economic Letters. 2007. V.96. P. 226-231.
Mika Meitz, Pentti Saikkonen. A note on the geometric ergodicity of a nonlinear AR-ARCH model // Statistics and Probability Letters. 2010. V. 80. P. 631-638.
Sean Mein, Richard Tweedie. Markov Chains and Stochastic Stability. Springer Verlag, 1993. 412 p.
Борисов, Конев В.В. О последовательном оценивании параметров дискретных процессов // Автоматика и телемеханика. 1977. № 10. С. 58-64
Дмитриенко А.А, Конев В.В. О последовательной классификации процессов авторегрессии с неизвестной дисперсией помех // Проблемы передачи информации. 1995. Т. 31. Вып. 4. С. 51-62.
Андерсон Т. Статистический анализ временных рядов. М.: Мир, 1976. 756 с. 10. ЛипцерР.Ш.,ШиряевА.Н. Теория мартингалов. М.: Наука, 1986. 512 с.
