Open Access Open Access  Restricted Access Subscription Access
Open Access Open Access Open Access  Restricted Access Restricted Access Subscription Access

Modeling the Mite Counts Having Overdispersion and Excess Values of Zero Using Zero-Inflated Generalized Poisson Regression


Affiliations
1 Biometry and Genetics Unit, University of Yuzuncu Yil, 65080 Van, Turkey
2 Usak University, Department of Econometry, 1 Eylul Kampusu, 64000, USAK, Turkey
     

   Subscribe/Renew Journal


The aim of this study was to apply for zero-inflated generalized Poisson regression in the modelling of mite counts that include excess values of zero and overdispersion. The results of ZIGP(μi φi ωi) , as mean regression, overdispersion and zero-inflated regression, were determined in three stages. It was obtained that 33.33% (120 observations) of the total numbers of mite taken as a dependent variable to model had zero values. The overdispersion parameter range was detected to be quite high. It was determined that zero-inflated data and overdispersion had an important effect on mite counts (P<0.01). The effects of region, month, year, varieties, temperature and humidity were found to be statistically significant on mite counts (P<0.01). The number of eggs found in harmful mites (Aculus schlechtendali) in the Starking variety was relatively higher than in the Golden variety. The results displayed that the differences among regions and varieties regarding the number of eggs found in harmful mites were statistically significant (P<0.01).

Keywords

Mite Counts, Overdispersion, Poisson Regression, Zero-Inflated Data.
Subscription Login to verify subscription
User
Notifications
Font Size


Abstract Views: 156

PDF Views: 1




  • Modeling the Mite Counts Having Overdispersion and Excess Values of Zero Using Zero-Inflated Generalized Poisson Regression

Abstract Views: 156  |  PDF Views: 1

Authors

Abdullah Yesilova
Biometry and Genetics Unit, University of Yuzuncu Yil, 65080 Van, Turkey
Baris Kaki
Usak University, Department of Econometry, 1 Eylul Kampusu, 64000, USAK, Turkey

Abstract


The aim of this study was to apply for zero-inflated generalized Poisson regression in the modelling of mite counts that include excess values of zero and overdispersion. The results of ZIGP(μi φi ωi) , as mean regression, overdispersion and zero-inflated regression, were determined in three stages. It was obtained that 33.33% (120 observations) of the total numbers of mite taken as a dependent variable to model had zero values. The overdispersion parameter range was detected to be quite high. It was determined that zero-inflated data and overdispersion had an important effect on mite counts (P<0.01). The effects of region, month, year, varieties, temperature and humidity were found to be statistically significant on mite counts (P<0.01). The number of eggs found in harmful mites (Aculus schlechtendali) in the Starking variety was relatively higher than in the Golden variety. The results displayed that the differences among regions and varieties regarding the number of eggs found in harmful mites were statistically significant (P<0.01).

Keywords


Mite Counts, Overdispersion, Poisson Regression, Zero-Inflated Data.