Recommenders: Add an experimentation example with AML

[miguelgfierro]

This article https://azure.microsoft.com/en-gb/blog/experimentation-using-azure-machine-learning/ from @danielsc can be used as a base

[loomlike]

I'm working on AML + hyperdrive for wide-deep, excluding auto-ml (don't think we --recommender-- can use automl). Basically I have two notebooks around 75% ready, one for wide-deep and the other one for aml+hyperdrive for wide-deep tuning

[yueguoguo]

FYI team I am also working on the Spark-based hyper tune which extends my current illustrative notebook here. Idea is to show how pipeline feature of AML can be used for tuning Spark models on Azure Databricks.

[miguelgfierro]

closing this since it has been solved on #448 and on #546