A class that extends the results of generalized linear models, glmssn
objects, for spatial stream networks by adding influence diagnostics and cross-validation
predictions to each observation.
Objects can be created by functions in the form residual(x),
where x is a glmssn-class object.
Objects of class influenceSSN contain 4 list items and have the exact same
structure as glmssn-class objects. A influenceSSN object retains
the corresponding SpatialStreamNetwork
object as the second list item. When residuals(x) is used for a glmssn object,
the data for which the model was fit is stored in point.data data.frame of the
observed points. This data.frame contains the response variable for the model,
and is appended by the following columns:
obsval ## The response value that was used to fit the model
_fit_
_resid_ ## The raw residuals
_resid.stand_ ## Standardized residuals; calculated by dividing the raw
## residuals by the corresponding standard errors
_resid.student_ ## Studentized residuals
_leverage_ ## Leverage
_CooksD_ ## Cook's D
_resid.crossv_ ## Cross-validation residuals
_CrossValPred_ ## Cross-validation predictions
_CrossValStdErr_ ## Estimated cross-validation standard errors.
Class "glmssn", directly.
Jay Ver Hoef support@SpatialStreamNetworks.com
residuals,glmssn