Evaluating model performance

Now we have created the model, let’s dive into the details of its performance.

When building machine learning models, it is very important to understand the model performance. You do this to make sure your model is useful and is not biased to one class over another and to make sure that the model is not under-trained or over-trained, which will mean the model is either not predicting classes correctly or is predicting only some instances and not others.

To address this problem, Redshift ML provides various objectives to measure the performance of the model. It is prudent that we test the model performance with the test dataset that we set aside in the previous section. This section explains how to review the Redshift ML objectives and also validate the model performance with our test data.

Redshift ML uses several objective methods to measure the predictive quality of machine learning models.

Checking the Redshift ML objectives

Figure...