hard20 pts · first answer countsmodel-monitoringforecastingmetrics

Comparing forecast errors fairly

A demand model’s MAPE "jumped" from 8% to 14% after a dashboard change. The new view includes 28-day-ahead forecasts, while the old one showed only 7-day-ahead. What is the right reading?

A.Error grows with horizon; compare like horizonsB.The model became much worse on all its forecastsC.MAPE cannot be computed for 28-day forecastsD.The 7-day forecasts must have leaked future data

Discussion 0

Join free to post in the discussion.

Comparing forecast errors fairly

Discussion 0

Related questions