Google Professional Machine Learning Engineer Question 4
Select 3Google Cloud PlatformYou are a machine learning engineer tasked with building a predictive model for customer churn using BigQuery ML. The dataset is already in BigQuery and contains millions of rows. Which of the following steps should you take to effectively develop the model using BigQuery ML?
- A
Preprocess the data using SQL-based transformations in BigQuery to handle missing values and normalize features.
- B
Export the dataset from BigQuery to a local environment and preprocess it using a Python library like pandas.
- C
Use the CREATE MODEL statement in BigQuery ML to train the model directly on the dataset.
- D
Manually split the dataset into training and testing sets before training the model.
- E
Enable feature engineering by defining feature transformations directly within the SQL query used for training.
Show answer and explanation
Correct answers: A, C, E
Explanation
BigQuery ML is designed to simplify the machine learning workflow by allowing you to preprocess data, engineer features, and train models directly within the BigQuery environment using SQL. There is no need to export data or manually split it for training. This approach is well-suited for handling large datasets efficiently.
- A. Correct.
Correct: BigQuery ML supports SQL-based preprocessing. Handling missing values and normalizing features directly in BigQuery is efficient for large datasets.
- B. Incorrect.
Incorrect: Exporting the dataset to a local environment is unnecessary, as BigQuery ML is designed to train models directly on data stored in BigQuery.
- C. Correct.
Correct: The CREATE MODEL statement is the standard way to train models in BigQuery ML, leveraging its built-in machine learning capabilities.
- D. Incorrect.
Incorrect: BigQuery ML automatically splits the dataset into training, evaluation, and testing sets, so manual splitting is not required.
- E. Correct.
Correct: BigQuery ML allows feature transformations to be applied via SQL queries, making it easier to engineer features during the training process.