Training thousands of forecasting models at scale using Azure Automated ML and ParallelRunStep

Author: Fatos Ismali

Business forecasting is becoming increasingly important in today’s dynamic and unpredictable world. Forecasting finds itself aplicable in a variety of industries and business scenarios — from predicting product demand in a marketing context to predicting energy consumption. In light of the recent Covid-19 pandemic many forecasting models have experienced what I call a “drift shock” that has meant that many of these models need to be retrained to take…