Google Professional Machine Learning Engineer Question 83
Select 4Google Cloud PlatformYou are tasked with building a time series forecasting model to predict monthly sales for the next year. Your company wants a solution that requires minimal custom code and leverages Google Cloud's managed services. After gathering the historical sales data, which of the following steps are necessary to create the forecasting model using AutoML in Vertex AI?
- A
Upload your historical sales data to a Google Cloud Storage bucket in CSV format.
- B
Preprocess your data manually to remove null values and scale all numeric features.
- C
Specify the target column and any feature columns when creating the AutoML training dataset.
- D
Configure the forecast horizon and granularity during the model training setup in Vertex AI.
- E
Deploy the trained model to an endpoint for batch or online predictions.
Show answer and explanation
Correct answers: A, C, D, E
Explanation
AutoML in Vertex AI simplifies the process of building forecasting models by automating many tasks such as preprocessing, model training, and hyperparameter tuning. However, users must follow specific steps, such as uploading data to Google Cloud Storage, defining target and feature columns, configuring the forecast horizon, and deploying the model for prediction purposes. Manual preprocessing is not required as AutoML can handle basic data preparation tasks.
- A. Correct.
Correct: AutoML in Vertex AI requires the data to be stored in a Google Cloud Storage bucket in a supported format (e.g., CSV) for training.
- B. Incorrect.
Incorrect: AutoML can handle basic preprocessing like handling missing values and scaling; manual preprocessing is not mandatory.
- C. Correct.
Correct: Defining the target column (e.g., sales) and feature columns (e.g., month, region) is a critical step when creating the training dataset in AutoML.
- D. Correct.
Correct: AutoML provides options to specify the forecast horizon (e.g., 12 months) and granularity (e.g., monthly) during model configuration.
- E. Correct.
Correct: After training, the AutoML model must be deployed to an endpoint to generate predictions for batch or online use cases.