A0702
Title: Robust estimation for bounded bivariate time series models of counts based on density power divergence
Authors: Minyoung Jo - Seoul National University (Korea, South) [presenting]
Sangyeol Lee - Seoul National University (Korea, South)
Abstract: Two types of bounded bivariate time series models are investigated, which are suitable for analyzing time series of counts with values within a finite range. They are a modified version of bounded bivariate integer-valued ARCH models and bounded bivariate INAR models. All these models are constructed based on the bivariate binomial distribution of Type II. The main focus is on providing a robust estimation method for these models. For this, the minimum density power divergence estimator (MDPDE) is employed as a robust estimator. To assess the performance of MDPDE and validate its effectiveness, both Monte Carlo simulations and real data analysis are conducted using monthly earthquake data in the United States. Findings collectively confirm the efficacy and suitability of the methods.
