Abstract
In this paper, we propose a novel index coefficient model for locally stationary time series data. Through a two-stage procedure, the estimation is derived and the asymptotic properties of the estimators are given. Moreover, we develop a test statistic for testing if the coefficient functions are time invariant, and derive the asymptotic distribution of the test statistic. A simulation study is conducted to assess the finite sample performances. Finally, a real data analysis is provided. The nice final sample simulation performances illustrate the efficiency of our proposed model and estimation.
| Translated title of the contribution | Statistical inference of time-varying single-index coefficient models for locally stationary time series |
|---|---|
| Original language | Chinese (Traditional) |
| Pages (from-to) | 1609-1630 |
| Number of pages | 22 |
| Journal | Scientia Sinica Mathematica |
| Volume | 50 |
| Issue number | 11 |
| DOIs | |
| Publication status | Published - Nov 2020 |
| Externally published | Yes |