I agree that UFO reports really need to be addressed in a more systematic and less biased fashion.
But I'm not convinced that there's any single methodological procedure that would enable us to identify all UFO reports. It's too heterogeneous a class, data is too fragmentary and the etiology is unknown (so different methods might be more applicable depending on the direction investigation takes).
Perhaps one way to approach it is to start out by characterizing the reports:
1. Single observer or multiple?
2. Single or multiple modalities? instrumental as well as visual? radar, cameras, etc.
3. If the observation was perceived to be extraordinary, what seemed extraordinary about it? configuration? behavior?
4. Any lingering physical effects observed on the surrounding environment?
5. Did the observers have any specialized expertise relevant to the observation?
6. Does the database contain reports of separate events that are similar in some respect? does whatever it is seem to be happening more than once?
And whatever more we can think of.
Then give the sightings values by how many of these kind of boxes they check off. It should be easy to code that kind of stuff into a computer database at least as yes-no. A more sophisticated version might code different kinds of extraordinary behavior, particular modalities, similar cases in particular respects, etc.
Doing that wouldn't really bring us a whole lot closer to actually identifying whatever the cause was, but it would highlight smaller classes of particular cases whose files need to be pulled and scrutinized.
I've called the 'tic-tac' sighting perhaps the best UFO reports that I'm personally aware of. (For whatever that's worth, it was just an observation about myself but it received lots of passionate opposition here on this board.) They would seem to me to check off five of the six boxes. (Lacking only lingering physical effects.)