Actions
Forum Thread Details
  • Replies 7 replies
  • Subscribers 595 subscribers
  • Views 1757 views
  • Users 0 members are here
Related

each time my post gets flagged for spam and 3 escort spams get on the main index, I melt a little

Jan Cumps

Motivational dip. I flagged 3 posts this week-end because they were for escorts, and they are still up. I wrote a nice blog post this week-end about an industrial sensor, related to a road test, and it's off to the spam queue.

image

Parents
  • Jan Cumps said:
    I flagged 3 posts this week-end because they were for escorts, and they are still up

    Are they? Could you provide a link so I can check that? As far as I can tell they've gone into the abuse queue on verint.


    Jan Cumps said:
    I wrote a nice blog post this week-end about an industrial sensor, related to a road test, and it's off to the spam queue.


    Yep... I had to raise the 'abusive link count' which didn't catch the escort spam which had just under the limit, and then other content gets caught for different reasons. Meanwhile there's still spam content that does get caught.

    I'm not a fan of how Verint's classifying the spam or its 'reputation system' (and neither are other customers of Verint). There are changes to it, but they're in the next version of the platform, and we're moving to that later this year.

    We need a way of whitelisting trusted users, and I don't have that yet (verint believes it should be the reputation system but it doesn't appear to work).

  • in reply to cstanton

    They're supposed to be famous for dodgy activity detection, it's one of their core businesses, I guess although on the one hand dodgy activity detection is supposed to pick out the anomalies in user activity (e.g. employees that act well for a long time, and then secretly start committing fraud), on the other hand for collaboration systems the system ought to trust more users that have not engaged in dodgy activity over repeated transactions (transaction being blog or comment writing in this context), so that manual whitelists would be an exception than a rule, but I guess they are still refining the major and the subtle differences between (say) financial systems, and more open collaboration systems. Interesting topic from a tech perspective, although of course it can get frustrating for users.

Reply
  • in reply to cstanton

    They're supposed to be famous for dodgy activity detection, it's one of their core businesses, I guess although on the one hand dodgy activity detection is supposed to pick out the anomalies in user activity (e.g. employees that act well for a long time, and then secretly start committing fraud), on the other hand for collaboration systems the system ought to trust more users that have not engaged in dodgy activity over repeated transactions (transaction being blog or comment writing in this context), so that manual whitelists would be an exception than a rule, but I guess they are still refining the major and the subtle differences between (say) financial systems, and more open collaboration systems. Interesting topic from a tech perspective, although of course it can get frustrating for users.

Children
No Data