A one class classification set-up for a set of rules acting as a model, where
- each input is a whole dataset
- model makes some decision within the dataset for each entry
- output is decisions made for each entry for that particular input
- reference standard judges whether the output is correct, calling it a "non-positive" if even one decision made for any entry is incorrect
Does this set-up make sense? Since I only have a say on positives, does using precision based on this reference standard(ground truth) make sense? I feel like there are gaps in this logic but can't pin-point where.