Article
Full entry |
PDF
(1.4 MB)
Feedback
PDF
(1.4 MB)
Feedback
Summary:
Some methods for approximating non-linear AR(1) processes by classical linear AR(1) models are proposed. The quality of approximation is studied in special non-linear AR(1) models by means of comparisons of quality of extrapolation and interpolation in the original models and in their approximations. It is assumed that the white noise has either rectangular or exponential distribution.
References:
[1] Anděl J.: On extrapolation in some non-linear AR(1) processes. Comm. Statist. – Theory Methods 26 (1997), 581–587 DOI 10.1080/03610929708831935 | MR 1436289
[2] Anděl J., Dupač V.: Extrapolations in non-linear autoregressive processes. Kybernetika 35 (1999), 383–389 MR 1704673
[3] Pemberton J.: Piecewise Constant Models for Univariate Time Series. Technical Report MCS-90-04, Department of Mathematics, University of Salford, Salford 1990
[4] Pemberton J.: Measuring nonlinearity in time series. In: Developments in Time Series Analysis (T. Subba Rao, ed.), Chapman and Hall, London 1993, pp. 230–240 MR 1292253 | Zbl 0880.62091
[5] Tong H.: Non-linear Time Series. Clarendon Press, Oxford 1990 Zbl 0835.62076
[6] Young P.: Time variable and state dependent modelling of non-stationary and nonlinear time series. In: Developments in Time Series Analysis (T. Subba Rao, ed.), Chapman and Hall, London 1993, pp. 374–413 Zbl 0880.62100
Search
Partner of
