$[16$pt$]$ Problem $5$: The "credit" dataset contains information on clients affiliated with Bank A$.$ The

dataset encompasses various variables, including

Status $-$ A categorical variable with two levels $($bad and good$)$ representing client's credit

score status $[$Target feature, "bad" is the target class$]$

Income $-$ Income in $$1000\text{'}$s

Limit $-$ Credit Limit

Rating $-$ Credit Rating

Cards $-$ Number of credit cards

Age $-$ Age in years

Own $-$ A factor with levels No and Yes, indicating whether the individual owns a home

Student $-$ A factor with levels No and Yes, indicating whether the individual is a student

Married $-$ A factor with levels No and Yes, indicating whether the individual is married

Region $-$ A factor with levels East, South, and West, indicating the individual's geographic

location

Balance $-$ Average credit card balance

The objective of this problem is to predict the status of clients using the remaining variables.

Using the parsnip package, build a predictive model utilizing the decision tree method. Use

the entire dataset and default parameters of the model. Using the model, make predictions

and generate a confusion matrix. Manually calculate the accuracy, sensitivity$,$ specificity, and

precision of the model.

Split the credit dataset into training and test sets by employing the stratified sampling

technique with a ratio of $70\%$ for training and $30\%$ for testing.

Construct a decision tree with the caret package: optimize the $"cp"$ by employing a bootstrap

resampling method $(15$ samples$).$ Choose the optimal hyperparameter value from the set

$\{0\mathrm{.}001,0\mathrm{.}01,0\mathrm{.}05,0\mathrm{.}1,0\mathrm{.}2\}$ based on the ROC criterion.

Display the outcomes of the hyperparameter tuning process by creating heatmaps that

illustrate sensitivity and specificity metrics.

Build the final model based on the optimal hyperparameter value obtained in part $(3).$ Using

the final model, make predictions for the training and testing sets.