Determine which is the best approach $($A or B$)$ for Question $1-5.$

A$.$ supervised learning B$.$ unsupervised learning

$1$ Separate customers into different groups based on their purchasing history.

$2$ Group all staff into different types based on their job performance.

$3$ Use sales revenue from the last quarter to predict sales revenue in next quarter.

$4$ Use outstanding balance and age to predict whether customers will default on the loan.

$5$ Separate credit card transactions into different types based on transaction time and location.

Determine whether the statements in Question $6-8$ is

A$.$ True B$.$ False

$6$ Decision Tree models cannot give us estimates of the probability of customer churn.

$7$ Support$-$Vector Machines $($SVMs$)$ can model non$-$linear relationships through kernel functions.

$8$ Adding more complexity to a model will generally increase its performance on the training set