Another one to add to it could be comment to post ratio, at face value this also may not mean much and even if it looks "healthy" an account could still just be commenting "nice post" or "!turdtoken" 10 times per day or just saying "thanks" to comments they receive while not giving a shit to consume and engage with other authors.
But since there are new accounts that do that every now and then, you do have to manage that.You could filter out comments that have a certain length. I'm doing that with a HiveSQL script for my @topcomment curation. Then you still have the accounts that have automated template replies which are long as well. I've excluded these accounts manually.
Maybe you could make a few indicators next to the rep score. For instance by using traffic light icons.
Not replying to comments on own posts --> Red
Good comment to post ratio --> Green
The more green, the better!
by the way: I designed KE to study behavior. I can demonstrate that there is a distinct difference in behavior if the KE of one account is 0.6 vs the one with 81 (both account holds 1000 HP). It is that difference I am after, and it is clear as daylight.
https://peakd.com/hive-133987/@azircon/governing-dynamics-20
what if your comments are AI generated.
I DV multiple hive accounts for that.
True.
I've seen quite some AI comments too indeed 😐
That is why we manually check comments we upvote with @topcomment.