‹ All AI-300 questions

AI-300 Practice Question 83: AutoML Forecasting Settings

Microsoft Machine Learning Operations Engineer Associate (AI-300) · Test 9 · question 3 of 10

Implement machine learning model lifecycle and operationshard

A retailer wants to predict daily demand for the next 14 days, separately for each of 300 stores, by using automated ML. The data has a date column and a store identifier column. Which configuration is correct?

Select one answer
Show answer and explanation

Answer:

  • C. A forecasting task with a 14-period horizon and the store column as series ID

Automated ML forecasting jobs need time_column_name and a forecast_horizon (here 14 periods). When the data holds several series, such as one per store, the identifier columns are given as time series identifiers so that AutoML models each series correctly (option C).

- A is wrong because a random split lets the model train on future dates, which leaks information and gives misleading validation scores. - B is wrong because demand is a numeric value over time. Predicting the store as a class does not forecast future demand. - D is wrong because the forecasting task requires the time column to be named, so ordering cannot be left to inference.

Reference: Set up AutoML for time-series forecasting

Practice Test 9 ›

Study guides for AI-300