Data Science Data Science Data Cleaning and Preparation Questions and Answers 3

Question 1

What is the main advantage of using K-Nearest Neighbors imputation over simple mean imputation for missing data?

Accepted Answer

It considers relationships between features to produce context-aware estimates

Answer

It is computationally faster on large datasets

Answer

It always produces integer values for categorical data

Answer

It requires no hyperparameter tuning

Question 2

When merging two datasets with different date formats (MM/DD/YYYY and YYYY-MM-DD), what should be done first?

Accepted Answer

Parse both columns into a standardized datetime format before merging

Answer

Convert all dates to Unix timestamps after merging

Answer

Delete rows where date formats do not match

Answer

Keep both formats and let the model handle the difference

Question 3

Which technique helps detect duplicate records that have slight variations in text fields such as customer names?

Accepted Answer

Fuzzy string matching using algorithms like Levenshtein distance

Answer

Exact string comparison after trimming whitespace

Answer

Sorting records alphabetically and visual inspection

Answer

Counting the frequency of each unique value

Question 4

What problem does multicollinearity cause during data preparation for regression models?

Accepted Answer

Unstable coefficient estimates that change dramatically with small data variations

Answer

Automatic removal of correlated features by the algorithm

Answer

Increased prediction accuracy on the training set

Answer

Reduced dimensionality of the feature space

Question 5

What is the purpose of applying a Box-Cox transformation during data preparation?

Accepted Answer

To make a skewed distribution more closely approximate a normal distribution

Answer

To remove all outliers from the dataset automatically

Answer

To convert categorical variables into numerical format

Answer

To reduce the number of features through dimensionality reduction

Question 6

When splitting data into training and test sets, why should stratified sampling be used for imbalanced classification problems?

Accepted Answer

To ensure both sets maintain the same class distribution as the original data

Answer

To guarantee equal numbers of samples in each class

Answer

To randomly shuffle the data before splitting

Answer

To remove minority class samples from the test set