Question $1$

What does the Receiver Operating Characteristic $($ROC$)$ curve plot in binary classification?

a$.$ Precision and recall values

b$.$ True positive rate against false positive rate

c$.$ True positive rate against true negative rate

d$.$ Sensitivity against specificity Explain your answer

Question $2$

What is the primary purpose of the gradient descent algorithm in machine learning?

a$.$ Maximize the cost function iteratively

b$.$ Minimize the cost function iteratively

c$.$ Find the large parameters for optimal solutions

d$.$ Maximize the number of iterations Explain your answer

Question $3$

Which describes the key difference between Lasso and Ridge regressions concerning feature coefficients?

a$.$ Ridge uses the L$1$ norm of the weight vector, Lasso uses half the square of the L$2$ norm

b$.$ Lasso adds a regularization term equal to the sum of the coefficients, while Ridge uses the L$1$ norm

c$.$ Lasso encourages feature selection by setting unimportant feature weights to zero; Ridge aims to keep all weights small

d$.$ Ridge outputs a sparse model with only a few nonzero weights; Lasso prioritizes increasing model complexity Explain your answer $.$

Question$4$

You are tasked with developing a binary classification ML model that classifies network traffic as either malicious or normal. You compute the confusion matrix below from your model. Using the values in the confusion matrix, calculate all five performance metrics listed below.

Prediction

Malicious normal

Malicious $599$

Actual Normal $11268$

a$.$ Accuracy

b$.$ Precision

c$.$ Recall

d$.$ Specificity

e$.$ F$1$ Score

Explain your answer by writing the correct equations and calculations for each metric. Provide your answers rounded to $2$ decimal places $($X$.$XX$).$