It’s just the spam filters. Sometimes there are false positives. We review the queues frequent and release such posts.
Hello Steven,
Thanks for answering!
It would really help if the spam filter would add some tags to the users that it’s marking as spammers, so that we could understand a bit of what it’s thinking.
Maybe then I could adjust my behavior so that it does not trigger again in the future.
Do you think that you could pass on this suggestion to the development team?
Best,
It would really help if the spam filter would add some tags to the users that it’s marking as spammers, so that we could understand a bit of what it’s thinking.
Nope. Never gonna happen. For two very good reasons:
1. If we told you why the system thought it was spam, then actual spammers would be able to work around it.
2. The system that figures out if it’s spam or not is an extremely complicated learning system named Akismet. Learning systems like this work based on some very weird math and other algorithms. In short, we have not got the faintest clue “why” it thought what it did. All we actually have to do is to correct it when it is wrong, and it learns, slightly, from that corrective action. We didn’t give it some set of fixed rules that are understandable by people, instead we gave it a bunch of actual spam and a bunch of non-spam and told it to figure out the difference. Which it did. So, we just correct it from time to time. Hey, it works pretty well, in the long run.
Hi Samuel,
Thanks for the reply!
All right, reason 1 is fair enough.
The black box argument for machine learning is open to question, but true, why bother getting some info if you’re not going to use it.
Cheers,
Gabriel