Indicating which of forecasting models at different aggregation levels has a better forecast quality

A method executed by a computer of performing forecasting, comprising:

• receiving a first data collection having data values at first intervals;
• performing statistical testing on the first data collection to derive a metric for indicating whether forecasting is to be performed at the first intervals of the first data collection or at second intervals based on aggregating the first intervals;
• in response to the metric indicating that forecasting is to be performed at the first intervals, applying a first forecasting model to produce forecast data at the first intervals;
• and
• in response to the metric indicating that forecasting is to be performed at the second intervals, applying a second forecasting model on a second data collection that has data values at the second intervals.

Instructions on a computer-usable storage medium that when executed cause a computer to:

• receive a first data collection having data values at first intervals;
• build a first forecasting model based on the first data collection;
• build a second forecasting model based on a second data collection having intervals aggregated from intervals of the first data collection, wherein the second forecasting model is at a different aggregation level than the first forecasting model;
• and
• compute at least one metric by performing at least one test based on at least one of the first and second data collections to indicate which of the first and second forecasting models has a better forecast quality.