Feature Engineering

Question 2

What is one-hot encoding?

Answer 2

Representing categorical variables as binary vectors.

Question 3

What is normalization?

Answer 3

Scaling data to a range, usually [0,1].

Question 4

What is standardization?

Answer 4

Transforming data to have mean 0 and variance 1.

Question 5

How to handle missing data?

Answer 5

Options: imputation, removal, or using algorithms that handle missing values.

Question 6

What is PCA?

Answer 6

Dimensionality reduction method finding orthogonal components maximizing variance.

Question 7

What is feature selection?

Answer 7

Choosing a subset of relevant features to improve model performance.

Question 8

What is multicollinearity?

Answer 8

When independent variables are highly correlated, affecting model stability.

Question 9

What is target encoding?

Answer 9

Encoding categories by replacing them with target variable statistics.

Question 10

What is binning?

Answer 10

Grouping continuous variables into discrete intervals.

Question 11

What is SMOTE?

Answer 11

Synthetic Minority Oversampling Technique for balancing class distribution.