Data Science Feature Engineering and Selection 2

Question 1

You apply a log transformation to a right-skewed feature containing zero values. What problem will this cause?

Accepted Answer

log(0) is undefined/negative infinity

Answer

It removes all outliers automatically

Answer

It converts the feature to categorical

Answer

It guarantees a normal distribution

Question 2

Which technique creates new features by multiplying or combining two existing features?

Accepted Answer

Interaction features

Answer

Standardization

Answer

Imputation

Answer

Label encoding

Question 3

When applying one-hot encoding to a feature with very high cardinality, the main drawback is:

Accepted Answer

A large, sparse feature space that increases dimensionality

Answer

Loss of all category information

Answer

Automatic target leakage

Answer

Conversion of numbers to text

Question 4

Target encoding replaces a category with what value?

Accepted Answer

The mean of the target for that category

Answer

A random integer

Answer

The category frequency rank

Answer

The standard deviation of the feature

Question 5

Which scaling method is most robust to outliers?

Accepted Answer

Robust scaling using median and IQR

Answer

Min-max scaling

Answer

Standardization (z-score)

Answer

Unit vector normalization

Question 6

Binning a continuous variable into discrete intervals is primarily used to:

Accepted Answer

Reduce the effect of minor observation errors and capture non-linearity

Answer

Increase the number of decimal places

Answer

Guarantee linear relationships

Answer

Remove the need for scaling entirely

Question 7

A polynomial feature of degree 2 on feature x adds which term?

Accepted Answer

x squared

Answer

log(x)

Answer

1/x

Answer

the square root of x