Google Professional Machine Learning Engineer Question 76
Select 2Google Cloud PlatformYou are a Machine Learning Engineer tasked with building a model to predict customer churn for a subscription-based business. The data includes both numerical and categorical features, such as customer age, subscription type, monthly usage, and whether they contacted support in the past month. You decide to use AutoML Tables in Google Cloud to build the model. Which of the following steps are required to ensure AutoML Tables works effectively for this tabular dataset?
- A
Upload the dataset to a Google Cloud Storage bucket and specify the target column when creating the AutoML Tables dataset.
- B
Manually scale all numerical features to a range between 0 and 1 before training the model.
- C
Ensure that categorical features are encoded as integers before uploading the dataset to AutoML Tables.
- D
Split the dataset into training, validation, and test sets during the dataset creation process in AutoML Tables.
- E
Remove any rows with missing values from the dataset before uploading it to AutoML Tables.
Show answer and explanation
Correct answers: A, D
Explanation
To effectively use AutoML Tables for tabular data, you must upload the dataset to a Google Cloud Storage bucket and specify the target column. AutoML Tables provides built-in preprocessing, including feature scaling, handling missing values, and encoding categorical features. Additionally, it is crucial to split the dataset into training, validation, and test sets during the dataset creation process to ensure proper model evaluation and tuning. Manual preprocessing tasks like scaling, encoding, or removing missing values are unnecessary as AutoML Tables automates these processes.
- A. Correct.
Correct. The dataset needs to be uploaded to a Google Cloud Storage bucket, and you must specify the target column so AutoML Tables knows what to predict.
- B. Incorrect.
Incorrect. AutoML Tables automatically handles feature scaling for numerical columns, so manual scaling is unnecessary.
- C. Incorrect.
Incorrect. AutoML Tables can handle raw categorical data without requiring manual integer encoding. It automatically encodes categorical features during preprocessing.
- D. Correct.
Correct. AutoML Tables allows you to split your dataset into training, validation, and test sets as part of the dataset creation process. This step is essential for proper model evaluation.
- E. Incorrect.
Incorrect. AutoML Tables can handle missing values and employs strategies like imputation to preprocess the data, so manual removal of rows with missing values is not required.