On the Optimal Adaptive Estimation in the Semiparametric Non-linear Autoregressive Time Series Model

  • So, Beong-Soo (Department of Statistics, Ewha Womens University, Seodaemun, Seoul 120-750)
  • Published : 1995.06.01

Abstract

We consider the problem of optimal adaptive estiamtion of the euclidean parameter vector $\theta$ of the univariate non-linerar autogressive time series model ${X_t}$ which is defined by the following system of stochastic difference equations ; $X_t = \sum^p_{i=1} \theta_i \cdot T_i(X_{t-1})+e_t, t=1, \cdots, n$, where $\theta$ is the unknown parameter vector which descrives the deterministic dynamics of the stochastic process ${X_t}$ and ${e_t}$ is the sequence of white noises with unknown density $f(\cdot)$. Under some general growth conditions on $T_i(\cdot)$ which guarantee ergodicity of the process, we construct a sequence of adaptive estimatros which is locally asymptotic minimax (LAM) efficient and also attains the least possible covariance matrix among all regular estimators for arbitrary symmetric density.

Keywords

References

  1. The Annals of Statistics v.11 Information and asymptotic efficiency in parametric-nonparametric models Begun,J.M.;Hall,W.J.;Huang,W.M.;Wellner,J.A.
  2. The Annals of Institute of Statistical Mathematics v.28 Adaptive estimates for autoregressive processes Beran,R.
  3. Journal of Time Series Analysis v.7 no.3 On estimating thresholds in autoregressive models Chan,K.S.;Tong,H.
  4. The Annals of Statistics v.6 On conditional least squares estimation for stochastic processes Klimko,L.A.;Nelson,P.I.
  5. The Annals of Statistics v.15 On adaptive estimation in stationary ARMA-models Kreiss,J.P.
  6. Contiguity of Probability Measures:Some Applications in Statistics Roussas,G.
  7. Journal of Statistical Planning and Infeence v.16 A note on the construction of asymptotically linear estimators Schick,A.
  8. Lecture notes in Statistics. 21 Threshold Models in Non-linear Time Series Analysis Tong,H.
  9. Non-linear Time Series Tong,H.