|
Activity Number:
|
519
- Innovations in Time Series Modeling
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 11, 2022 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
|
Abstract #322886
|
|
|
Title:
|
A Generalization to time-varying Integer GARCH Model
|
|
Author(s):
|
Isuru Panduka Ratnayake* and V. A. Samaranayake
|
|
Companies:
|
Kansas University Medical Center and Missouri University of Science and Technology
|
|
Keywords:
|
Count Data;
Integer Valued Time Series;
INGARCH;
Time Varying Parameters
|
|
Abstract:
|
This paper presents a time-varying Poisson process is proposed to model count data time series with serial dependence. The model assumes that the intensity of the Poisson process evolves according to a generalized conditional autoregressive heteroskedastic (GARCH) model with time varying GARCH parameters. The proposed generalization to time-varying Poisson Integer GARCH (G-tvINGARCH) model is a generalization of the Integer GARCH (INGARCH) model proposed by Ferland, Latour, and Oraichi in 2006. The proposed model builds on these previous formulations by incorporating GARCH parameter to vary over time and be driven by an exogenous variable. Such a parameterization would be more appropriate to capture time varying conditional heteroskedastic behavior in count data. The Maximum Likelihood Estimation (MLE) approach is presented as possible estimation method. The results of a Monte-Carlo simulation study indicate that the MLE methods works well in producing relatively accurate parameter estimates. An empirical study is presented to illustrate the performance of the proposed model in a real-life setting.
|
Authors who are presenting talks have a * after their name.
