Takes a fitted addreg.smooth
object produced by addreg.smooth
and plots
the component smooth functions that make it up, on the scale of the linear predictor,
for specified values of the other covariates.
# S3 method for addreg.smooth
plot(x, type = c("response", "link"), at = data.frame(),
knotlines = TRUE, nobs = 1000, ...)
a fitted addreg.smooth
object as produced by addreg.smooth
.
the type of prediction required. Note that, unlike predict.addreg.smooth
,
"terms"
is not a valid option. Also, because addreg.smooth
only applies
identity-link models, "response"
and "link"
will have the same results ---
they are included for consistency.
a data frame containing the values at which the prediction should be evaluated. The columns
must contain the covariates in the model, and several rows may be provided (in which case, multiple
lines are drawn on the same plot). Cannot be missing or NULL
.
logical; if vertical lines should be drawn on the plot to indicate the locations of the knots for B-spline terms.
the number of points which should be used to create the curve. These are placed evenly along the range of the observed covariate values from the original model.
other graphics parameters to pass on to plotting commands (note: some will not work).
The function simply generates plots.
For each smooth covariate in the model of x
, predict.addreg.smooth
is used to obtain predicted values for the range of that covariate, with the other
covariates remaining fixed at their values given in at
. Several rows may be provided
in at
, in which case, one curve is drawn for each, and they are coloured using
rainbow(nrow(at))
. If the model contains a single smooth covariate and no other
covariates, at
may be provided as an empty data frame, data.frame()
.
# NOT RUN {
## For an example, see example(addreg.smooth)
# }
Run the code above in your browser using DataLab