A0542
Title: Inference of signal variance in time series for mean stationarity test
Authors: Hon Kiu To - The Chinese University of Hong Kong (Hong Kong) [presenting]
Kin Wai Chan - The Chinese University of Hong Kong (Hong Kong)
Abstract: Inference of mean structure is an important problem in time series analysis. Various tests have been developed to test for different mean structures, including but not limited to, the presence of structural break(s), and parametric mean structures. Many of them are designed under specific mean structures, and may potentially lose power upon violation of such structures. We propose a new mean stationarity test built around the signal variance. The proposed test can detect the non-constancy of the mean function under serial dependence. It is shown to have promising power in detecting hardly noticeable periodic structures. The proposal is further generalized to test for smooth mean structures and the relevant structural changes in time series. A real-data application on global land surface temperature data is presented.
