We used the AUC value to measure the accuracy of our binary classification model. For multivalue classification models, AWS uses the macroaverage F1 score to evaluate the predictive accuracy of a multiclass metric. A larger F1 score indicates better predictive accuracy for a regression model; AWS uses the root mean square error (RMSE) metric. The smaller the value of the RMSE, the better the accuracy of the model. 

A detailed discussion of ML concepts is beyond the scope of this book. The aim of this recipe was to get you started with Amazon ML using AWS CLI APIs, and to familiarize you with a few ML terms that you can explore further. You can follow the reference links or other books on ML to learn more and experiment with the concepts further. I have also added links to some datasets in the See also section that you can use for your experiments.