Scholarlink

This work presents an application of the Pseudo-Variance Quasi-Maximum Likelihood Estimators (PVQMLE) for estimating Integer-valued Autoregressive (INAR) models. We apply this methodology to two diverse datasets: software download and animal skin lesions datasets. The approach is showed to outperform other quasi-likelihood methods, offering superior goodness-of-fit performance. Additionally, it enables the selection of thinning and innovation distributions within the INAR model framework. By proposing this new estimation technique, we enhance the accuracy and flexibility of modeling integer-valued time series data, thereby facilitating more robust analysis in various fields. This research contributes to advancing statistical methodologies for analyzing discrete-valued time series data, offering valuable insights into complex phenomena represented by such datasets.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

On the Estimation of INAR Models with Pseudo-Variance Quasi-Maximum Likelihood

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Mirko Armillotta,
  • Paolo Gorgi

摘要

This work presents an application of the Pseudo-Variance Quasi-Maximum Likelihood Estimators (PVQMLE) for estimating Integer-valued Autoregressive (INAR) models. We apply this methodology to two diverse datasets: software download and animal skin lesions datasets. The approach is showed to outperform other quasi-likelihood methods, offering superior goodness-of-fit performance. Additionally, it enables the selection of thinning and innovation distributions within the INAR model framework. By proposing this new estimation technique, we enhance the accuracy and flexibility of modeling integer-valued time series data, thereby facilitating more robust analysis in various fields. This research contributes to advancing statistical methodologies for analyzing discrete-valued time series data, offering valuable insights into complex phenomena represented by such datasets.