Bivariate Poisson Generalized Lindley Distributions and the Associated BINAR(1) Processes

Authors

  • Irshad Muhammed Rasheed CUSAT
  • Veena Dcruz
  • Maya Radhakumari University of Kerala
  • Soobhug Ashwinee Devi University of Mauritius
  • Noushad Mamode Khan

DOI:

https://doi.org/10.17713/ajs.v53i1.1551

Abstract

This paper proposes new bivariate distributions based on the Poisson generalized Lindley distribution as marginal. These models include the basic bivariate Poisson generalized Lindley (BPGL) and the Sarmanov-based bivariate Poisson generalized Lindley (SPGL) distributions. Subsequently, we introduce the BPGL and SPGL distributions as joint innovation distributions in a novel bivariate first-order integer-valued autoregressive process (BINAR(1)) based on binomial thinning. The model parameters in the BPGL and SPGL distributions are estimated using the method of maximum likelihood (ML) while we apply the conditional maximum likelihood (CML) for the BINAR(1) process. We conduct some simulation experiments to assess the small and large sample performances. Further, we implement the new BINAR(1)s to the Pittsburgh crime series data and they show better fitting criteria than other competing BINAR(1) models in the literature.

Downloads

Published

2024-01-15

How to Cite

Muhammed Rasheed, I., Dcruz, V., Radhakumari, M., Ashwinee Devi, S., & Mamode Khan, N. (2024). Bivariate Poisson Generalized Lindley Distributions and the Associated BINAR(1) Processes. Austrian Journal of Statistics, 53(1), 1–28. https://doi.org/10.17713/ajs.v53i1.1551