Question

A data scientist is preparing a customer churn dataset in a Spark DataFrame. The categorical column 'preferred_channel' has about 8% missing values, and the business confirms these are missing at random with no meaningful 'unknown' category. The scientist wants to impute the missing values before one-hot encoding, while avoiding data leakage. Which approach is most appropriate?

Accepted Answer

Compute the most frequent value (mode) of 'preferred_channel' on the training set only, then use that value to fill missing entries in both the training and test sets.

Answer

Compute the mode of 'preferred_channel' on the full dataset before splitting, then fill missing values and split afterward.

Answer

Fill missing values in 'preferred_channel' with the numeric mean of its one-hot encoded columns computed across the entire dataset.

Answer

Drop all rows where 'preferred_channel' is missing to guarantee no imputation bias is introduced.