# R/plot.select.parfm.R In parfm: Parametric Frailty Models

#### Documented in plot.select.parfm

```################################################################################
#  Plot of objects of class 'select.parfm'                                     #
################################################################################
#                                                                              #
#                                                                              #
#                                                                              #
#   Date: December 22, 2011                                                    #
#   Last modification on: December  2, 2016                                    #
################################################################################

plot.select.parfm <- function(x,
mar=c(2.5, 2, 1.5, .5),
ty = 'b',
...){
par(mfrow=c(1, 3))

### --- AIC --- ###
par(mar=mar)
plot(0,0, ty="n", xlab="", ylab="", main="AIC", xaxt="n",
ylim=c(min(x\$AIC, na.rm=TRUE) *  .9975,
max(x\$AIC, na.rm=TRUE) * 1.0025),
xlim=c(.5, ncol(x\$AIC) + .5), cex.lab=1.5)
abline(v=1:ncol(x\$AIC), col="grey")

mtext(c(none="No",
gamma="Ga",
ingau="IG",
possta="PS",
lognor="LN")[colnames(x\$AIC)],

for (i in 1:nrow(x\$AIC)) points(
(1:ncol(x\$AIC)), x\$AIC[i, ],
pch = 19 + i, cex = 1.5, ty = ty, bg = i)

### --- names --- ###
par(mar=mar)
plot(0:2, 0:2, xaxt = "n", yaxt = "n", bty = "n", ann = FALSE,
ty = "n")

legend("top", #c(.3, 1.7), c(1, 1.75),
title = 'Baseline',
c(exponential = "exponential",
weibull = "Weibull",
inweibull = "inverse Weibull",
gompertz = "Gompertz",
loglogistic = "loglogistic",
lognormal = "logNormal",
logskewnormal = "logSkewNormal")[rownames(x\$AIC)],
pch = {if(ty == 'l') NULL else 19 + 1:nrow(x\$AIC)},
pt.bg = 1:nrow(x\$AIC),
bg = "white", bty = "n", lty = ifelse(ty == 'p', 0, 1),
ncol = 1, cex = 1.5, xjust = .5)

legend("bottom", #c(0, 2), c(.25, 1), yjust=1,
title = 'Frailty distribution',
mapply(paste,
c(none="No",
gamma="Ga",
ingau="IG",
possta="PS",
lognor="LN")[colnames(x\$AIC)],
c(none="no frailty",
gamma="gamma",
ingau="inverse Gaussian",
possta="positive stable",
lognor="lognormal")[colnames(x\$AIC)],
sep=" = "),
bg="white", bty="n",
ncol=1, cex=1.5, xjust=.5)
### --- end names --- ###

### --- BIC --- ###
par(mar=mar)
plot(0,0, ty="n", xlab="", ylab="", main="BIC", xaxt="n",
ylim=c(min(x\$BIC, na.rm=TRUE) *  .9975,
max(x\$BIC, na.rm=TRUE) * 1.0025),
xlim=c(.5, ncol(x\$BIC) + .5), cex.lab=1.5)
abline(v=1:ncol(x\$BIC), col="grey")

mtext(c(none="No",
gamma="Ga",
ingau="IG",
possta="PS",
lognor="LN")[colnames(x\$BIC)],