Add documentation samples for binary classifiers using the dynamic API

[sfilipi]

Similar to PR #1361 we need code samples to reference from the new MLContext catalog extensions for all binary learners.

Reference to the sample in PR #1361 and add samples for the following binary trainers.
Cross reference those samples from the respective extensions on the catalog files.

Trainers:
SGD
LogisticRegression
FFM
FastTree
LightGBM

[jwood803]

@sfilipi Not sure if this is exactly what is needed, but I have this example using Logistic Regression if it helps.

[sfilipi]

@jwood803 that example is perfect. DO you have interest/time to contribute it to the repo here, under the Microsoft.M.Samples project, and reference it from the LogisticRegression catalog? Unfortunately, it would have to be a different dataset, maybe the "adult" one, as we don't have permission to use the Titanic data.

[jwood803]

@sfilipi I'll be glad to put it under the samples project. Just to make sure, I should use this dataset, correct?

[sfilipi]

Anyone of the datasets in this folder should do. We can't use datasets outside of this repo...

Look at this PR, for a guide on how to link the sample back to the code it illustrates.

Thanks for taking this @jwood803 ! Let's leave the issue open, since there are another 4 trainer estimators to illustrate.

[sfilipi]

This is now complete. Progress was also tracked through: #2522