How can I update the code below to:

a$)$ Manually specify the cost complexity paramente with a specific value. for example, create a regular grid with only one value for the complexity parameter?

b $)$ Prune the tree. Create a regular or random grid with some values for the cost complexity parameter, re$-$run the hyperparameter tuning step, and investigate the performance of the different hyperparameter values?

c$)$ Select the best one according to for example ROC$\_$AUC or Accuracy?

d$)$ Create the training set, validation, and test prediction using the pruned decision tree.

e$)$ Visualize the pruned decision tree?

dataset $<-$ read$\_$csv$("$Census$.$csv$")$

str$($dataset$)$

dataset$Income $<-$ as$.$factor$($dataset$Income$)$

dataset$Education.num $<-$ as$.$factor$($dataset$Education.num$)$

dataset$Race $<-$ as$.$factor$($dataset$Race$)$

dataset$Sex $<-$ as$.$factor$($dataset$Sex$)$

dataset $=$ select$($dataset$,-3)$

dataset$Income $<-$ factor$($dataset$Income, levels $=$ c$("$Low$",$"High"$))$

str$($dataset$)$

summary$($dataset$)$

plot$\_$intro$($dataset$)$

set.seed$(1)$

data$\_$split $<-$ initial$\_$split$($data $=$ dataset,

prop $=0\mathrm{.}8,$

strata $=$ Income$)$

data$\_$train $<-$ training$($data$\_$split$)$

data$\_$test $<-$ testing$($data$\_$split$)$

target$\_$var $<-$ "Income"

model$\_$form $<-$ Income ~ Age $+$ Education.num $+$ Race $+$ Sex $+$ Hours.per.week

positive$\_$class $<-$ 'Low'

model$\_$recipe $<-$ recipe$($formula $=$ model$\_$form, data $=$ data$\_$train$)\%>\%$

step$\_$novel$($all$\_$nominal$\_$predictors$())\%>\%$

step$\_$unknown$($all$\_$predictors$(),-$all$\_$numeric$())\%>\%$

step$\_$dummy$($all$\_$nominal$\_$predictors$())\%>\%$

step$\_$zv$($all$\_$predictors$())$

class$\_$tree$\_$model $<-$

decision$\_$tree$($cost$\_$complexity $=$ tune$())\%>\%$

set$\_$engine$("$rpart$")\%>\%$

set$\_$mode$("$classification$")$

class$\_$tree$\_$workflow $<-$ workflow$()\%>\%$

add$\_$recipe$($model$\_$recipe$)\%>\%$

add$\_$model$($class$\_$tree$\_$model$)$

all$\_$models $<-$ workflow$\_$set$($

preproc $=$ list$($model$\_$recipe$),$

models $=$ list$($class$\_$tree$\_$rpart $=$ class$\_$tree$\_$model$)$

$)$

set.seed$(1)$

cv$\_$folds $<-$ vfold$\_$cv$($data$\_$train, v$=5)$

control $<-$ control$\_$resamples$($save$\_$workflow $=$ TRUE,

save$\_$pred $=$ TRUE,

event$\_$level $=$ "second"$)$

classification$\_$metrics $<-$ metric$\_$set$($yardstick::accuracy,

yardstick::kap,

yardstick::roc$\_$auc,

yardstick::mn$\_$log$\_$loss,

yardstick::sens$,$

yardstick::spec,

yardstick::f$\_$meas,

yardstick::precision,

yardstick::recall$)$

tGrid$\_$random $<-$ grid$\_$random$($cost$\_$complexity$(),$

size $=50)$

models$\_$trained $<-$ all$\_$models $\%>\%$

workflow$\_$map$("$tune$\_$grid",

resamples $=$ cv$\_$folds,

grid $=$ tGrid$\_$random,

metrics $=$ classification$\_$metrics,

verbose $=$ TRUE,

control $=$ control$)$