rsm.diag.plots {marg}

Diagnostic Plots for Regression-Scale Models

Description

Generates diagnostic plots for a regression-scale model using different types of residuals, Cook's distance and the leverages.

Usage

rsm.diag.plots(rsmfit, rsmdiag = NULL, weighting = NULL, 
               which = NULL, subset = NULL, iden = FALSE, 
               labels = NULL, ret = FALSE, ...)
## S3 method for class 'rsm'
plot(x, ...)

Arguments

rsmfit, x

a rsm object, i.e. the result of a call to rsm.

rsmdiag

the object returned by a call to rsm.diag containing the regression diagnostics for the regression-scale model defined by rsmfit. If not supplied, this object is created by rsm.diag.plots and returned upon request (if ret = TRUE).

weighting

character string; defines the weight matrix that should be used in the calculation of the residuals and diagnostics. Possible choices are "observed", "score", "deviance" and "max"; see Jorgensen (1984) for their definition. Will only be used if the rsmdiag argument is missing.

which

which plot to print. Admissible values are 2 to 7 corresponding to the choices in the menu below.

subset

subset of data used in the original rsm fit: should be the same than the subset option used in the call to rsm which generated rsmfit. Needed only if the subset option was used in the call to rsm.

iden

logical argument. If TRUE, the user will be prompted after the plots are drawn. A positive integer will select a plot and invoke identify() on that plot. After exiting identify(), the user is again prompted, this loop continuing until the user responds to the prompt with 0. If iden is FALSE (default) the user cannot interact with the plots.

labels

a vector of labels for use with identify() if iden is TRUE. If it is not supplied, then the labels are derived from rsmfit.

ret

logical argument indicating if rsmdiag should be returned; the default is FALSE.

...

additional arguments such as graphical parameters.

Details

The diagnostics required for the plots are calculated by rsm.diag. These are then used to produce the plots on the current graphics device.

A menu lists all the plots that can be produced. They may be one or all of the following:

 Make a plot selection (or 0 to exit)

1: All
2: Response residuals against fitted values
3: Deviance residuals against fitted values
4: QQ-plot of deviance residuals
5: Normal QQ-plot of r* residuals
6: Cook statistic against h/(1-h)
7: Cook statistic against observation number

Selection:
  

In the normal scores plots, the dotted line represents the expected line if the residuals are normally distributed, i.e. it is the line with intercept 0 and slope 1.

In general, when plotting Cook's distance against the standardized leverages, there will be two dotted lines on the plot. The horizontal line is at 8/(n-2p), where n is the number of observations and p is the number of estimated parameters. Points above this line may be points with high influence on the model. The vertical line is at 2p/(n-2p) and points to the right of this line have high leverage compared to the variance of the raw residual at that point. If all points are below the horizontal line or to the left of the vertical line then the line is not shown.

Use of iden = TRUE is encouraged for proper exploration of these plots as a guide to how well the model fits the data and whether certain observations have an unduly large effect on parameter estimates.

Value

If ret is TRUE then the value of rsmdiag is returned, otherwise there is no returned value.

Side Effects

The current device is cleared. If iden = TRUE, interactive identification of points is enabled. All screens are closed, but not cleared, on termination of the function.

Acknowledgments

This function is based on A. J. Canty's function glm.diag.plots contained in the package boot.

References

Davison, A. C. and Snell, E. J. (1991) Residuals and diagnostics. In Statistical Theory and Modelling: In Honour of Sir David Cox (eds. D. V. Hinkley, N. Reid, and E. J. Snell), 83–106. London: Chapman & Hall, London.

Davison, A. C. and Tsai, C.-L. (1992) Regression model diagnostics. Int. Stat. Rev., 60, 337–353.

Jorgensen, B. (1984) The Delta Algorithm and GLIM. Int. Stat. Rev., 52, 283–300.

See Also

rsm.diag, rsm.object, identify

Examples

## Sea Level Data
data(venice)
attach(venice)
Year <- 1:51/51
c11 <- cos(2*pi*1:51/11) ; s11 <- sin(2*pi*1:51/11)
c19 <- cos(2*pi*1:51/18.62) ; s19 <- sin(2*pi*1:51/18.62)
venice.rsm <- rsm(sea ~ Year + I(Year^2) + c11 + s11 + c19 + s19, 
                  family = extreme)
## Not run: 
rsm.diag.plots(venice.rsm, which = 3)

## End(Not run)
## or
## Not run: 
plot(venice.rsm)

## End(Not run)
## menu-driven
##
rsm.diag.plots(venice.rsm, which = 5, las = 1)
## normal QQ-plot of r* residuals 
## Not run: 
rsm.diag.plots(venice.rsm, which = 7, iden = T, labels = paste(1931:1981))

## End(Not run)
## year 1932 highly influential
detach()