Plots distributed lags curves from the modTempEff
fit.
1 2 3 4 
x 
object of class 
which 
Which DL curve should be plotted? for cold, heat or both of them (default). 
add 
logical; if 
new 
logical indicating if a new device should be opened. If 
var.bayes 
logical indicating if the 'Bayesian' rather than the frequentist standard errors should be employed to compute the pointwise confidence intervals to be plotted 
delta.rr 
logical indicating if the DL curves should be plotted on the log scale or as per cent change in relative risk, i.e. 100*(exp(.)1). 
level 
the selected confidence level of the pointwise confidence intervals to be plotted 
updown 
logical; if 
col.shade 
the color of the shaded area representing the pointwise confidence intervals. If 
leg 
the possible legends to be set on the 
... 
additional arguments, such as 
Takes a fitted "modTempEff"
object produced by tempeff()
and plots the
DL curves for cold and heat effect with relevant pointwise confidence intervals.
plot.modTempEff
also works with objects with fixed (not estimated) breakpoint, namely
fits returned by
tempeff(.., fcontrol=fit.control(it.max=0))
.
Note add=TRUE
makes sense (and works) only for a single (cold or heat) DL curve to be superimposed to an existing plot.
The function simply plots the required estimated DL curve. If the fitted model includes only a smooth term for
the long term trend, plot.modTempEff
draws it.
Vito Muggeo
tempeff
1 2 3 4 5 6 7 8 9 10 11 12  ## Not run:
#obj is an object returned by tempeff()
#plots DL curves for cold and heat with 95% pointwise CI
# using frequentist standard errors
plot(obj)
#plots the estimated DL curve only for heat with 90% pointwise CI
# using bayesian standard errors
plot(obj, "heat", var.bayes=TRUE, level=.90)
## End(Not run)

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.
All documentation is copyright its authors; we didn't write any of that.