ecision trees have great expressive power: they can represent any decision boundary, by carving up the

space to a fine enough granularity. Which of the other methods we$$ve studied are similarly expressive?

For each, give a brief explanation.

$($a$)$ Linear classifiers.

$($b$)$ Support vector machines with a quadratic

kernel.

$($c$)$ Nearest neighbor classifiers.

$($d$)$ Classifiers based on Gaussian generative models.decision