Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 425 - Nonparametric Methods for Dependent Data
Type: Contributed
Date/Time: Wednesday, August 10, 2022 : 10:30 AM to 12:20 PM
Sponsor: Section on Nonparametric Statistics
Abstract #320845
Title: Tilted Density Function Estimation: On the Bickel-Rosenblatt Test for Nonlinear Autoregressive Models
Author(s): Yan Su and Shuxia Sun* and Fuxia Cheng
Companies: Wright State University and Wright State University and Illinois State University
Keywords: Bickel-Rosenblatt test; goodness-of-fit; logistic smooth transition; nonlinear autoregressive model; stochastic process,; tilted density function estimator

This paper considers the goodness-of-fit test on the error distribution in nonlinear autoregressive time series models. We propose a new type of the Bickel-Rosenblatt test by use of the tilted density estimator based on residuals from the fitted model and show that, under appropriate conditions, the corresponding test statistic has an asymptotic normal distribution. Moreover, our simulation studies indicate that this new type of test is more powerful than the one based on the conventional kernel error density estimator. A real data example is also presented to demonstrate the practical implementation of this goodness-of-fit test.

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

Back to the full JSM 2022 program