Robust recursive estimation in nonlinear time series |
| |
| Authors: | T. Cipra |
| |
| Affiliation: | Dept. of Statistics , Charles University of Prague , Sokolovská 83, 186 00, Czech RepublicPrague 8 |
| |
| Abstract: | Aase (1983) has dealt with recursive estimation in nonlinear time series of autoregressive type including its asymptotic properties. This contribution modifies the results for the case of nonlinear time series with outliers using the principle of M-estimation from robust statistics. Strong consistency of the robust recursive estimates is preserved under corresponding assumptions. Several types of such estimates are compared by means of a numerical simulation. |
| |
| Keywords: | Robust recursive estimation nonlinear time series outliers strong consistency |
|
|
