The R Project SVN R

Rev

Rev 74265 | Rev 78962 | Go to most recent revision | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 74265 Rev 75334
Line 34... Line 34...
34 
\alias{covratio}
 34 
\alias{covratio}
35 
\alias{cooks.distance}
 35 
\alias{cooks.distance}
36 
\alias{cooks.distance.lm}
 36 
\alias{cooks.distance.lm}
37 
\alias{cooks.distance.glm}
 37 
\alias{cooks.distance.glm}
38 
\usage{
 38 
\usage{
39 
influence.measures(model)
 39 
influence.measures(model, infl = influence(model))
40 
 
 40 
 
41 
rstandard(model, \dots)
 41 
rstandard(model, \dots)
42 
\method{rstandard}{lm}(model, infl = lm.influence(model, do.coef = FALSE),
 42 
\method{rstandard}{lm}(model, infl = lm.influence(model, do.coef = FALSE),
43 
          sd = sqrt(deviance(model)/df.residual(model)),
 43 
          sd = sqrt(deviance(model)/df.residual(model)),
44 
          type = c("sd.1", "predictive"), \dots)
 44 
          type = c("sd.1", "predictive"), \dots)
Line 115... Line 115...
115 
  \code{rstudent} give the standardized and Studentized residuals
 115 
  \code{rstudent} give the standardized and Studentized residuals
116 
  respectively. (These re-normalize the residuals to have unit variance,
 116 
  respectively. (These re-normalize the residuals to have unit variance,
117 
  using an overall and leave-one-out measure of the error variance
 117 
  using an overall and leave-one-out measure of the error variance
118 
  respectively.)
 118 
  respectively.)
119 
 
 119 
 
- 
 
 120 
  Note that for \emph{multivariate} \code{lm()} models (of class
- 
 
 121 
  \code{"mlm"}), these functions return 3d arrays instead of matrices,
- 
 
 122 
  or matrices instead of vectors.
- 
 
 123 
 
120 
  Values for generalized linear models are approximations, as described
 124 
  Values for generalized linear models are approximations, as described
121 
  in Williams (1987) (except that Cook's distances are scaled as
 125 
  in Williams (1987) (except that Cook's distances are scaled as
122 
  \eqn{F} rather than as chi-square values).  The approximations can be
 126 
  \eqn{F} rather than as chi-square values).  The approximations can be
123 
  poor when some cases have large influence.
 127 
  poor when some cases have large influence.
124 
 
 128