Please report false positives!

in #spam6 years ago

learning_from_this_one.png

Learning more every day

I had my first false positive today on this comment by a user who definitely did say "beautiful picture" and "thanks for sharing" and 37 other similar phrases in the last 100 comments, but is clearly and consistently engaging with the post.

I've been really happy with the fact that I've literally seen no false positives until now. To be fair, I tested and looked at the data a LOT before I started commenting, so my algorithm definitely looks for the right stuff, but I did expect to see more false positives.

That said, I'm sure there's some out there, I just haven't seen them so I'm asking for your help. If you come across my reply to something that looks legitimate, please take a moment to scan the last few comments on the account and if it doesn't look too much like spam, reply to my comment to let me know.

These false positives are extremely useful as I can use the data of the entire account to find other patterns to exclude these legitimate users. Today I have matching patterns to find these comments and now I'm building my exclusionary patterns to avoid commenting on legit users.

While my true hope is for my algorithm to be accurate already, to the extent it isn't I hope you can help me find any false positives and help with my continued learning and growth!

Coin Marketplace

STEEM 0.31
TRX 0.12
JST 0.034
BTC 64742.01
ETH 3172.49
USDT 1.00
SBD 4.10