Add result reporting/parsing. #22
Description
Mainly to help users wade through the massive output this module can produce, and to help them make conclusions on how to proceed.
In particular, the following features need to be added:
- Reporting the best performance per-replicate
- Visual hyper-parameter sampling (both per replicate and across the entire study; this is needed to help decide what "best" protocol should be used)
- Visualized performance over time, both for the "objective" and other specified metrics requested by the user (i.e. feature importance).
- Converting the database into a spreadsheet for easier distribution/management.
I'm open to further suggestions as well; this just needs to be done before we can really hope for any usage outside of people already familiar with programming ML tools.