Google Professional Machine Learning Engineer Question 63
Select 3Google Cloud PlatformYou are tasked with building a machine learning model for predicting customer churn using AutoML Tables in Google Cloud. The dataset consists of 500,000 rows with features such as customer demographic details, account history, and usage patterns. Before importing the data into AutoML Tables, what steps should you take to ensure the dataset is properly prepared for training?
- A
Ensure all features have the same scale by normalizing numeric columns.
- B
Remove all categorical features since AutoML Tables does not support them.
- C
Manually label your target variable if it is missing or incomplete.
- D
Split your dataset into training, validation, and test sets after importing it into AutoML Tables.
- E
Check for and handle missing values in the dataset before importing into AutoML Tables.
Show answer and explanation
Correct answers: A, C, E
Explanation
Preparing data for AutoML requires ensuring features are correctly scaled, the target variable is properly labeled, and the dataset is clean and free of missing values. While AutoML Tables automates many tasks, such as dataset splitting, the initial data preparation remains a critical step for ensuring high-quality training and predictions.
- A. Correct.
Correct. Normalizing numeric columns can help ensure better performance of the model during training.
- B. Incorrect.
Incorrect. AutoML Tables supports categorical features and can automatically handle them by encoding them as needed.
- C. Correct.
Correct. The target variable must be complete and accurately labeled for AutoML Tables to train a predictive model.
- D. Incorrect.
Incorrect. AutoML Tables automatically splits the data into training, validation, and test sets, so this does not need to be done manually.
- E. Correct.
Correct. Handling missing values before importing ensures the dataset is clean, which can improve model performance.