This seems to be incredibly simple and user-friendly.
However, transparency would be helpful regarding the inner workings of the bot and what its judgment really implies. Or maybe all I need, really, would be some intuition on the degree of confidence. Maybe an "orginality" score instead of a digital outcome would be helpful.
Hi @gustave.
It is funny that you say that, because initially that was the direction I was going too.
Technically this is doable, and the algorithm is such that it is well suited for that.
As of now I am working towards "perfecting" the accuracy, although it isn't there yet it is pretty good right now.
That being said, if I use a score based system, people who use quotes might come out with a low score. I am trying to avoid this, because currently, even if you have quotes, as long as the rest of your material is original this bot will mark it as original.
It is a good thought though, and something I will investigate once I clean up everything else a bit more.
As far as how it works, I do not want to reveal too much and allow for abuse, but it is essential a fragmentation based algorithm. Feel free to try it with plagiarized articles, and let me know how it goes!
Thank you for your input! This kind of input truly helps me improve the bot.
I see the point - a bad rating for quoting a lot seems counterproductive.