Databricks Machine Learning Associate Question 521
Select 3You are working on a dataset in Databricks that contains missing values in a numerical column called age. You aim to impute these missing values to make the dataset suitable for machine learning modeling. Given the following scenarios, which imputation technique(s) would be most appropriate?
- A
Use the mean value of the
agecolumn if the data is symmetrically distributed. - B
Use the median value of the
agecolumn if the data contains significant outliers. - C
Use the mode value of the
agecolumn if the data is categorical. - D
Use the mean value of the
agecolumn if the data is highly skewed. - E
Use the median value of the
agecolumn if the data is symmetrically distributed.
Show answer and explanation
Correct answers: A, B, C
Explanation
Imputation techniques depend on the distribution and type of the data. The mean is suitable for symmetrically distributed numerical data, the median is more robust for data with outliers, and the mode is best for categorical data. Choosing the correct imputation strategy ensures that the dataset is accurately prepared for machine learning models.
- A. Correct.
Using the mean value is appropriate when the data is symmetrically distributed, as it best represents the central tendency in such cases.
- B. Correct.
Using the median value is effective when the data contains significant outliers, as the median is resistant to the influence of extreme values.
- C. Correct.
The mode value is suitable for categorical data since it represents the most frequently occurring value in the dataset.
- D. Incorrect.
Using the mean value for highly skewed data is not ideal, as the mean is affected by extreme values and may not accurately represent the central tendency.
- E. Incorrect.
Using the median for symmetrically distributed data is unnecessary, as the mean would better represent the central tendency in this case.