As a wizard-user I can inspect prediction results, in order to prioritize maintenance work and learn about my project's dynamics

[fkoehne]

This is meant to provide an easy access to the results for users that are not familiar with machine learning terminology and methods.
As a wizard-user, I can intuitively see whether the model is precise enough to serve as a decision foundation at all, i.e. by contextualizing it's RMSE-Values.
Further, I can see the most important factors for achieving this result (as provided by h2o for non-ensemble models).
Most importantly I receive a list of the files contained and their change probability rating in descending order (i.e. top priorities for maintenance work regarding change probability)

[fkoehne]

This is meant as an introduction or overview, we will further need an expert perspective with more details for those that are familiar with machine learning terminology and methods.