Mixed models applied to the litterfall productivity
A methodology is developed to adjust a linear mixed model incorporating a variancecovariance structure starting from a the classical linear regression model following Zuur strategy, Zuur et al. (2009). Different diagnostic plots of residuals and likelihood ratio tests are used to choose fixed and rand...
| Autor Principal: | Zea, José F. |
|---|---|
| Otros Autores: | Murcia, Miguel A., Poveda, Fidel E. |
| Formato: | info:eu-repo/semantics/article |
| Idioma: | spa |
| Publicado: |
Universidad Santo Tomás
2014
|
| Acceso en línea: |
http://revistas.usta.edu.co/index.php/estadistica/article/view/1482 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: |
A methodology is developed to adjust a linear mixed model incorporating a variancecovariance structure starting from a the classical linear regression model following Zuur strategy, Zuur et al. (2009). Different diagnostic plots of residuals and likelihood ratio tests are used to choose fixed and random effects and to find an appropriate covariance structure to capture the heteroskedasticity and serial correlation of model residuals. Ecological data from Murcia (2013) research around the river basin Pamplonita (Norte de Santander, Colombia) are used in this paper, different measurements are taken in the high Andean and sub-Andean forests. The weight of litter fallen from 40 collectors located in two forest was quantified for 12 months. This data are analyzed using mixed models. |
|---|