So… i'm planning to create a website to rival the likes of MyAnimeList and Anime-Planet.
Am I Insane? Yes. Yes I am.So anyway, the basic motivation behind this is my annoyance with the way people score things. Most people score things way too damn high, with 7's equaling "okay/mediocre" and everything below being "a waste of time". Not everybody scores like this, though… some people score more evenly. And some people have their entirely own way of scoring things. Which creates a serious problem when you try to average everybody's score to create a meaningful average for an anime. It just doesn't work.How anime will be ranked on my site is different. Users don't give out numerical scores. Instead, they put anime into two personal lists: "I Recommend" and "I Don't Recommend" and ranks them in order from strongest to least strongest. The "machine" then derives a score based on which list it's on and its position on that list. This, I hope, will completely eliminate bias generated by the act of scoring. The only "bias" will be the user's opinion of the anime, which is perfect because we're trying to come to score that most accurately describes the average opinion of the anime.Ambitious perhaps, but I'm completely serious about it. I've already got a host and a url. The site isn't anything yet, but here's the URL:http://www.ar-machine.net
Didn't notice that NSFW, I just saw a link, got excited, and clicked it.
std::string AnimeRecommendationMachine() { return "Boku no Pico"; }IT'S A TRAP!
Sounds cool, If you make it- I'll bring my 356 Anime buddies with me.
Make sure It lists hentai.Yuri Hentai.Updated the site a bit. Added a work-in-progress stylesheet and made it to where you're no longer greeted with hentai immediately upon entry.
Earlier, there wasn't a "random image" link… the homepage just displayed a random image. lol
Thank fuck for that, indeed. But according to my maths, my site's served up somewhere between 439 and 2748 images… LOL.
But now the images are semi-SFW. They're all clothed and not displaying sexual acts or fluids, anyways.i don't think this will work. the most popular anime will get to the top while hundreds of gems will be left in the dust. also odds are there won't be as many positive votes as there are negative votes. i mean look at Youtube's system. the majority of videos are always highly positive because people looking up something they're interested in means they already think positive about it!
now i know you got your idea, but here's mine (just for fun). I think it would be more effective to track individuals' likes and dislikes, and match them with similar individuals, than suggest to similar individuals popular likes amongst individuals similar to them. (hope that's not too confusing xD)std::string AnimeRecommendationMachine() { return "Boku no Pico"; }I don't think people need to see Thirteen year old faggots having sex, Rob.