AWS Certified Machine Learning – Specialty — Question 278

A machine learning (ML) specialist at a retail company must build a system to forecast the daily sales for one of the company's stores. The company provided the ML specialist with sales data for this store from the past 10 years. The historical dataset includes the total amount of sales on each day for the store. Approximately 10% of the days in the historical dataset are missing sales data.

The ML specialist builds a forecasting model based on the historical dataset. The specialist discovers that the model does not meet the performance standards that the company requires.

Which action will MOST likely improve the performance for the forecasting model?

Answer options

• A. Aggregate sales from stores in the same geographic area.
• B. Apply smoothing to correct for seasonal variation.
• C. Change the forecast frequency from daily to weekly.
• D. Replace missing values in the dataset by using linear interpolation.

Correct answer: D

Explanation

Missing data in time-series forecasting can significantly degrade model performance, and replacing these gaps using linear interpolation preserves the sequential structure of the daily sales data. Aggregating regional stores (Option A) or changing the forecast frequency (Option C) alters the business requirements of predicting daily sales for that specific store. Applying smoothing (Option B) removes seasonal variations rather than addressing the underlying issue of missing data points.