MS-DS Master of Data science Feature Engineering 3

Question 1

Which technique addresses class imbalance at the feature level by creating synthetic minority samples?

Accepted Answer

SMOTE (Synthetic Minority Over-sampling Technique)

Answer

PCA whitening

Answer

Weight of Evidence encoding

Answer

Variance Inflation Factor reduction

Question 2

What is Weight of Evidence (WoE) encoding primarily used for?

Accepted Answer

Encoding categorical features for binary classification by measuring the strength of each category in predicting the target

Answer

Encoding continuous features as categorical bins

Answer

Weighting features by their inverse frequency in the dataset

Answer

Normalizing sparse text features in NLP pipelines

Question 3

What is the purpose of the Variance Inflation Factor (VIF) in feature engineering?

Accepted Answer

Detecting multicollinearity among predictor features

Answer

Measuring how much a feature's variance changes after scaling

Answer

Estimating how much a feature inflates model variance compared to a null model

Answer

Selecting the subset of features with the highest explained variance

Question 4

In natural language processing feature engineering, what does TF-IDF measure?

Accepted Answer

The importance of a term in a document relative to how common it is across the entire corpus

Answer

The total frequency of a term across all documents in a corpus

Answer

The similarity between two documents based on shared terms

Answer

The number of unique tokens in each document normalized by document length

Question 5

Which feature engineering approach is most appropriate for handling a date column (e.g., '2024-03-15') in a structured dataset?

Accepted Answer

Extracting sub-features such as year, month, day of week, and is_weekend

Answer

Label encoding the full date string into a unique integer

Answer

Dropping the column because dates are not numeric

Answer

Applying log transformation to the Unix timestamp

Question 6

What does a Box-Cox transformation accomplish in feature engineering?

Accepted Answer

It maps a positive-valued feature to a more normal distribution by finding an optimal power parameter λ

Answer

It converts categorical variables into continuous probability scores

Answer

It centers and scales features to zero mean and unit variance

Answer

It applies a piecewise linear transformation to cap outliers at specified percentiles

Question 7

Which of the following best describes 'feature hashing' (hashing trick)?

Accepted Answer

Mapping high-cardinality or text features into a fixed-size numerical vector using a hash function

Answer

Applying a cryptographic hash to anonymize sensitive feature values

Answer

Generating a unique integer ID for each feature to reduce memory usage

Answer

Encoding features by their MD5 hash to guarantee collision-free representation