Task $1$

Apply suitable data pre$-$processing operations on the data:

$1.$ Identify in your report $($not just in your R code$)$ what pre$-$processing operations

you performed.

$(5$ points$)$

$2.$ Produce R code in RStudio that implements these pre$-$processing operations.

$(10$ points$)$

Task $2$

$1.$ Produce R code for implementing the classifiers in the pre$-$processed dataset.

$(10$ points$)$

$2.$ Produce rule bases for your classifiers. Save these rule bases into your pdf

report document and present them under the heading $$Task $2.$

$(5$ points$)$

Task $3$

In this task we will use $10-$fold cross validation $($cv$)$ with $3$ repeats to obtain an

accurate assessment of a classifier$$s accuracy. The cv strategy takes into account

variation of data across a dataset and is a more reliable assessment of accuracy.

Once all the folds $(30)$ are in place you will carry out an F test to compare the $3$ types

of classifiers.

Collect the accuracy across the classifiers into $3$ vectors from the $30$ different folds.

Carry out an F test in R $($One$-$way ANOVA $|$ When and How to Use It $($With Examples$)$

$($scribbr$.$com$)$ on the $3$ vectors.

Results of the F test may reveal that the $3$ classifiers have significantly different

accuracies at a $95\%$ level of confidence or that they have no significant difference. If

the F test reveals a significant difference, then the next step would be to determine

which one of them has the highest accuracy. This can be done with the help of the

Tukey post hoc test, which is covered from the same website as the one given for

the one$-$way ANOVA.

$1.$ Produce the R code for this task.

$(25$ points$)$

$2.$ Copy and paste the ANOVA table into your pdf report.

$(5$ points$)$

$3.$ Identify, which $($if any$)$ classifier has the highest accuracy if it happens that the

F test reveals that the $3$ are significantly different $($use the Tukey test here$).$

Noe that this part is not needed if all $3$ classifiers are flagged by the F test as

having the same level of accuracy.

$(5$ points$)$

For this submission, your R code should be clearly commented, and each task should be labeled. Add

print statements to your code to clearly demonstrate it working. Your code should be able to run

directly in RStudio without any compilation errors or warnings. This task should not require any user

input.