We first need to establish a baseline. Amazon ML quantile binning transformation is Amazon Machine Learning preferred method of dealing with non-linearities in a dataset. Let's establish how a simple linear model performs with Amazon's default recipe. We will create a baseline score using the usual AWS console tools. This time, we choose to create a data source from Redshift and not from S3. Fill the information as shown in the next screenshot, and click on Test access to check your access, and at the same time, make Amazon ML create the necessary IAM role. Finally, click on Verify once you're done:
Amazon ML handles ...
