automatically create empirical and kde model

[nandaloo]

Models include a link to a 'data' model, the emp_model attribute. They are a immediate representation of the data, in comparision to other, advanced statistical models. Currently this model must be created manually for each model. This is tedious and error prone. Hence, it should be automated.

If a data model is queried it can come with additional, optional information on how to derive this model on the fly. Hence, if the model does not exist, and it provides these additional information, it should be created and stored.

• define optional arguments
• detect missing model when querying and if these arguments are available
• automatically create model

In particular a facitily is require to automatically create data models. Currently, I can see two suitable model types: KDE and emp. I think both functionality exists somewhere already.

• implement automatic creation of empirical model from data
• implement automatic creation of KDE model from data

With the completion of this issue data model are now handled fully automatically. Hence, there should be a detection when the model has to be recreated, even if it exists:

• implement update checks (time stamp based is easiest)
• implement creation time stamp for all models by default

Before starting out you need to verify that this works at all with the fronend. Check the places where empirical models are queried and imagine to add the additional optional arguments. Can you see any issues?

• done