Tom_Ph|||pp| @end|ng |rom np@@gov
Sat Jul 4 01:20:12 CEST 2020
Poisson and Negative Binomial distributions are defined for counts of whole numbers. Even if the original dependent variables were counts, by definition PCA creates continuous numeric variates. Therefore, Poisson or Negative Binomial are not appropriate for variates from PCA.
I suggest that you shift this over to R-sig-ecology, and give background on your experimental or sampling design, and questions of interest. At least for now, this is not a question about Mixed Effects, although once the design is specified, it may require a mixed effects analysis.
Inventory and Monitoring Program
National Park Service
"To do science is to search for repeated patterns, not simply to accumulate facts..." --Robert MacArthur 1972, Geographical Ecology
"Statistical methods of analysis are intended to aid the interpretation of data that are subject to appreciable haphazard variability" --Cox & Hinkley 1974; Theoretical Statistics
Respect science, respect nature, respect each other.
From: R-sig-mixed-models <r-sig-mixed-models-bounces using r-project.org> On Behalf Of Saâd HANANE
Sent: Tuesday, June 30, 2020 1:08 PM
To: r-sig-mixed-models using r-project.org
Subject: [EXTERNAL] [R-sig-ME] PCA-GLMM
This email has been received from outside of DOI - Use caution before clicking on links, opening attachments, or responding.
I have performed a PCA analysis which highlighted two axes PC1 and PC2. I want to perform a GLMM analysis considering the PC1 axis as a dependent variable (variable to explain) with Poisson or negative binomial family, but it was not possible since the values of this axis are negative. In such a case what should I do?
I also need to perform glmmPQL to consider spatial autocorrelation. LME could be a solution (instead of GLMM)? If yes, how could I perform glmmPQL?
Thank you for your precious help.
Saâd Hanane, PhD
Service d'Ecologie, de Biodiversité et de Conservation des Sols Centre de Recherche Forestière Chariae Omar Ibn Al Khattab, BP 763, Rabat-Agdal/Maroc.
[[alternative HTML version deleted]]
R-sig-mixed-models using r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
More information about the R-sig-mixed-models