Title: Asymptotic properties of mildly explosive processes with locally stationary disturbance
Authors: Junichi Hirukawa - Niigata University (Japan) [presenting]
Sangyeol Lee - Seoul National University (Korea, South)
Abstract: For a class of locally stationary process, we consider the problem of testing ARMA model against other non-stationary ARMA model. When testing the problem, we use linear serial rank statistics and contiguity of LeCam's notion. If the null hypothesis is white noise, then, under the null and the alternative, the asymptotic normality of the proposed statistics is established by using the locally asymptotic normality (LAN). We incorporate the locally stationary phenomena in the testing problem.