AWS Certified Machine Learning – Specialty — Question 343

A clothing company is experimenting with different colors and materials for its products. The company stores the entire sales history of all its products in Amazon S3. The company is using custom-built exponential smoothing (ETS) models to forecast demand for its current products. The company needs to forecast the demand for a new product variation that the company will launch soon.

Which solution will meet these requirements?

Answer options

• A. Train a custom ETS model.
• B. Train an Amazon SageMaker DeepAR model.
• C. Train an Amazon SageMaker К-means clustering model.
• D. Train a custom XGBoost model.

Correct answer: B

Explanation

Amazon SageMaker DeepAR is an algorithm designed for training a forecasting model across multiple related time series, which makes it highly effective for predicting the demand of new products (cold-start scenario) by leveraging historical data from similar items. Custom ETS models cannot easily generalize to new products that lack individual historical sales data. K-means is an unsupervised clustering algorithm and is not suitable for time-series demand forecasting.