Poisson inverse Gaussian Liu-type estimator: a comparative study with Monte Carlo simulation

Işılar M., BULUT Y. M.

Soft Computing, 2026 (SCI-Expanded, Scopus) identifier

  • Yayın Türü: Makale / Tam Makale
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1007/s00500-026-11293-z
  • Dergi Adı: Soft Computing
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, zbMATH
  • Anahtar Kelimeler: Count data, Multicollinearity, Overdispersion, Poisson inverse Gaussian regression
  • Eskişehir Osmangazi Üniversitesi Adresli: Evet

Özet

The Poisson regression model (PRM), which is commonly used to model count data, cannot be applied to over- or under-dispersed data sets. The Negative Binomial regression model (NBRM) is preferred as an alternative to the Poisson regression model for modeling overdispersed count data. As the degree of overdispersion increases, the performance of NBRM may decrease. In this situation, the Poisson Inverse Gaussian regression model (PIGRM) can serve as an alternative to PRM and NBRM for modeling highly overdispersed count data. We introduce and evaluate Liu-type estimators as a novel and efficient method to mitigate multicollinearity effects in the PIGRM, thereby improving the stability and accuracy of parameter estimates. A simulation study is carried out to examine the performance of the proposed estimators. Also, a real data example is provided to support the simulation results.