Databricks Machine Learning Associate Question 2
Select 2You are building a machine learning pipeline in Databricks to predict customer churn. The dataset you are working with contains missing values and categorical features. You plan to use Databricks' AutoML to automate model selection and hyperparameter tuning. Before running AutoML, what steps should you take to ensure your dataset is ready for AutoML?
- A
Ensure all missing values are handled, as AutoML does not handle missing data automatically.
- B
Convert categorical features into numerical representations using one-hot encoding or similar techniques.
- C
Split the dataset into training, validation, and test sets before running AutoML.
- D
Ensure the dataset is stored as a Delta table or a Spark DataFrame.
- E
Manually scale numerical features to standardize their range before running AutoML.
Show answer and explanation
Correct answers: A, D
Explanation
When using Databricks AutoML, it is important to prepare your dataset by handling missing values and ensuring the data is in the correct format (Delta table or Spark DataFrame). AutoML handles other preprocessing steps such as feature scaling, categorical encoding, and dataset splitting automatically.
- A. Correct.
Correct. Databricks AutoML does not automatically handle missing values, so you must ensure missing values are either imputed or removed before running AutoML.
- B. Incorrect.
Incorrect. Databricks AutoML can handle categorical features automatically by applying encoding techniques, so manual encoding is not required.
- C. Incorrect.
Incorrect. AutoML will handle dataset splitting into training, validation, and test sets automatically, so manual splitting is unnecessary.
- D. Correct.
Correct. Databricks AutoML requires input data to be in a Delta table or Spark DataFrame format to function properly.
- E. Incorrect.
Incorrect. Databricks AutoML automatically handles feature scaling as part of its preprocessing pipeline, so manual scaling is not needed.