Model performance

Let’s use Redshift SQL to compute a confusion matrix to evaluate the performance of the classification model. Using a confusion matrix, you can identify true positives, true negatives, false positives, and false negatives, based on which various statistical measures such as accuracy, precision, recall, sensitivity, specificity, and finally, F1 score are calculated. You can read more about the concept of the confusion matrix here: https://en.wikipedia.org/wiki/Confusion_matrix.

The following query uses a WITH clause, which implements a common table expression in Redshift. This query has the following three parts:

• The first part is about the SELECT statement within the WITH clause, where we predict customer churn and save it in memory. This dataset is named infer_data.
• The second part, which is below the first SELECT statement, reads infer_data and builds the confusion matrix, and these details are stored in memory in a dataset called confusionmatrix...