As discussed, we want to gather statistics to aid our understanding of what exactly is going on in the data. This will be useful in further analyses of the methods and their behavior.
Some considerations:
at the very least, we'd like to see the EGR/class probabilities charts
it would be redundant to run this analysis on every experiment, so we'll probably need a dedicated dataset analysis script
As discussed, we want to gather statistics to aid our understanding of what exactly is going on in the data. This will be useful in further analyses of the methods and their behavior.
Some considerations: