#skewness

Question 1: Single choice

You build a linear model to predict house prices. The prices are all positive and strongly right-skewed: most houses are cheap, a few are very expensive. What is the usual first preprocessing step for the target?

Fit the model to $\log(\text{price})$ and exponentiate its predictions1
Replace the prices by one-hot encoded price bands2
Delete every house priced above the 95th percentile, treating them as outliers3
Min-max scale the prices to the interval $[0, 1]$ 4

Timed test