Online Program Home
My Program

Abstract Details

Activity Number: 614 - Statistical Methods for Longitudinal and Other Dependent Data
Type: Contributed
Date/Time: Thursday, August 1, 2019 : 8:30 AM to 10:20 AM
Sponsor: Section on Nonparametric Statistics
Abstract #307203 Presentation
Title: Comprehensive Simultaneous Inference on Trend-Cycle Model
Author(s): Sayar Karmakar* and Wei Biao Wu
Companies: University of Florida and University of Chicago
Keywords: Simultaneous inference; Confidence band; Cyclostationary; Trend; Non-stationary

Cyclostationary process is a special type of non-stationary processes showing a periodic pattern. We analyze the popular trend-cycle model in presence of a very general non-stationary error and build simultaneous inference on the trend function post the estimation of the period length and the periodic component. Our simultaneous confidence bands are comprehensive in the sense it does not suffer from the boundary problem and thus it could allow a more general ARMA-GARCH type of noise processes that essentially requires larger bandwidth to estimate. We also discuss how to jointly study multiple such similar series and test whether the periods are synchronized or not. We conclude by providing some simulations and analysis of temperature anomaly data.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program