MLS-C01 · Question #20
An online reseller has a large, multi-column dataset with one column missing 30% of its data. A Machine Learning Specialist believes that certain columns in the dataset could be used to reconstruct…
The correct answer is C. Multiple imputation. Multiple imputation is the correct choice when the analyst believes other columns contain information useful for reconstructing the missing values. It builds a predictive model (often regression) using the observed columns to impute the missing column multiple times, creating…
Question
An online reseller has a large, multi-column dataset with one column missing 30% of its data. A Machine Learning Specialist believes that certain columns in the dataset could be used to reconstruct the missing data. Which reconstruction approach should the Specialist use to preserve the integrity of the dataset?
Options
- AListwise deletion
- BLast observation carried forward
- CMultiple imputation
- DMean substitution
How the community answered
(54 responses)- A6% (3)
- B2% (1)
- C81% (44)
- D11% (6)
Explanation
Multiple imputation is the correct choice when the analyst believes other columns contain information useful for reconstructing the missing values. It builds a predictive model (often regression) using the observed columns to impute the missing column multiple times, creating several plausible complete datasets and pooling results - preserving the relationships between variables and the distributional integrity of the data. Listwise deletion (A) removes all rows with any missing value, losing 30% of the data and potentially introducing bias. Last observation carried forward (B) is a time-series technique inappropriate here. Mean substitution (D) is simple but ignores inter-variable relationships and artificially reduces variance.
Topics
Community Discussion
No community discussion yet for this question.