Understanding your model accuracy
Some items might be more important than others in a dataset. In retail forecasting, 20% of the sales often accounts for 80% of the revenue, so you might want to ensure you have a good forecast accuracy for your top-moving items (as the others might have a very small share of the total sales most of the time). In every use case, optimizing accuracy for your critical items is important: if your dataset includes several segments of items, properly identifying them will allow you to adjust your forecast strategy.
In this section, we are going to dive deeper into the forecast results of the first predictor you trained in Chapter 4, Training a Predictor with AutoML. In Chapter 6, Generating New Forecasts, I concatenated all the forecast files associated with the AutoML model trained in Chapter 4, Training a Predictor with AutoML. Click on the following link to download my Excel file and follow along with my analysis: