Deep diving into forecasting model metrics
When training a forecasting model, Amazon Forecast will use different machine learning metrics to measure how good a given model and parameters are at predicting future values of a time series. In this section, we are going to detail what these metrics are, why they are important, and how Amazon Forecast uses them.
In Chapter 4, Training a Predictor with AutoML, you trained your first predictor and obtained the following results:
Using the dropdown on the top right of this section, you can select the algorithm for which you want to see the results. The results you obtained from your first predictor should be similar to the following ones:
At the top of this table, you can see several metric names, such as wQL[0.1], wQL[0.5], wQL[0.9], WAPE, and RMSE. How are these metrics computed? How...