Improve clarity of Reference Documentation for METRICS

[CESARDELATORRE]

I'm finding that most metrics have very little explanation in the reference.
Other times, it just tells what's the formula for that metric.

For example:

https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.data.regressionmetrics.rootmeansquarederror?view=ml-dotnet

https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.data.multiclassclassificationmetrics.logloss?view=ml-dotnet

From a developer point of view, what we need to explain is:

• Very short summary of what is that metric (not a formula. The formular is nice to have but not the fundamental thing to understand the metric)
• What to look for with that metric: "The closer to 1.00 the better?", viceversa?, basically, when the model is better or worse when looking at that metric.

Other samples where there's very little explanation about the metric:

https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.data.multiclassclassificationmetrics.microaccuracy?view=ml-dotnet

https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.data.regressionmetrics.rsquared?view=ml-dotnet

https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.data.regressionmetrics.meanabsoluteerror?view=ml-dotnet