Please note that neither myself nor other members of our group are exactly statisticians or high-level r coders.
ASREML R AR1 AT HOW TO
Can anyone refer me to a recent discussion regarding this type of analysis? Or can they offer advice on how to construct a mixed effects models that accounts for spatial correlation in nlme or lmer? The consensus at the time seemed to be that it wasn’t straight forward to do this in either lmer or nlme.Īfter spending a few hours searching, it’s not totally clear to me whether there has been any progress on addressing this. Several years ago a question was posted asking if an open-source alternative to asreml-r was available that could be used to construct a two-dimensional spatial model with error structure in both direction. So they are reluctant to either pay for asreml-r or consider using. My research group also prefers to rely on nlme and lmer for the majority of it’s analytical needs. Unfortunately, asreml-r is expensive and difficult to learn.
It is relatively easy to use the asreml package to specify a spatial model (i.e. Like many things in fixed models, they do not always carry over simly to mixed models.In the past, I have used asreml-r to account for spatial auto-correlation in agricultural field trials that were laid out in a “row and range” design.
So, it is convenient to talk about residuals in a loose sense, but when looking at their atributes, we need to think about their correlation structure and exactly where they fit into the mixed model. The IID residuals as the deviations from the fitted model. Now you could add to the model a random term 'units' which are assumed IID and are residuals in that sense though fitted in the G structure.Īnd of course you could fit the A1 x AR1 'residuals' as random effects and So, yes under a typical AR1 x AR1 model, the 'residuals are correlated. Nevertheless, most of us were first introduced to residuals as IID Normal variables in a simle fixed model setting. I am unaware of any rule that says residuals should be independent any more than that they should have equal variance. Which is the residual from the fixed model. The residual can be regarded as Zu+e (wth variance R+ZGZ') In ASReml, I have used the term residual loosely to refer to the random term in the model that has variance structure R. The marginal residual is $\vec$) sometimes referred to as drop-one residuals and are the residuals obtained from predicting the observation ignoring the actual observation (estimated using the out() term in ASReml as a fixed effect. Haslett and Haslett (I say a draft paper, 2006, H\&H, hopefully it is published now) review residualsĪnd identify three basic kinds: marginal, model and conditional residuals.
Hopefully others will contribute to this discussion. I'm glad to see someone adding a bit of interest to this forum.
0 kommentar(er)
