Databricks Machine Learning Associate Question 189
Single answerYou are working on a machine learning pipeline in Databricks. Your dataset contains a categorical feature 'Product_Category' with the values ['Electronics', 'Clothing', 'Furniture']. You decide to apply one-hot encoding to this feature. After encoding, which of the following statements about the resulting dataset is correct?
- A
The 'Product_Category' column will be replaced with three new columns, each representing one category.
- B
The 'Product_Category' column will remain in the dataset, with numerical values replacing the category names.
- C
The resulting encoded columns will contain values in the range [0, 1], where 1 indicates the presence of that category.
- D
The one-hot encoding process is irreversible, and the original categorical column cannot be reconstructed from the encoded columns.
Show answer and explanation
Correct answer: A
Explanation
One-hot encoding is a technique to transform categorical features into a numerical format suitable for machine learning algorithms. It replaces the categorical column with multiple binary columns, where each column corresponds to a unique category in the original feature. Each binary column contains a 1 if the record belongs to that category and 0 otherwise. This process is reversible as long as the mapping between the categories and their respective columns is retained.
- A. Correct.
Correct. One-hot encoding replaces the categorical column with new binary columns, one for each unique value in the original column.
- B. Incorrect.
Incorrect. One-hot encoding replaces the original column entirely; it does not retain the column with numerical values.
- C. Incorrect.
Incorrect. The values in the resulting columns are binary (0 or 1), not within a range of [0, 1].
- D. Incorrect.
Incorrect. One-hot encoding is reversible as long as the original mapping of categories to columns is preserved.