11.6 MODEL EVALUATION FOR ESTIMATION
We can use the regression equation to make predictions (estimates) of sales per visit. For example, consider Customer 1 in the training set, who goes 333 days between purchases and who does not have a store credit card (Credit Card = 0). Plugging these values into the regression equation, we obtain:
That is, using the regression model, we would estimate the average sales per visit for this customer to be
So, this customer spends $56.10 more than expected, given his or her days between visits and credit card status.
The typical size of the prediction error is given by the statistic s, the standard error of the estimate.