For my previous post zemanta generated following tags: South Park, Television program, Fridays, Park, Television, BitTorrent, Kids and Teens, Arts
Only half of the tags are related to my post. I couldn’t find the point for BitTorrent and my Friday Night post? Maybe for South Park? Zemanta may think most important thing in my post is South Park and added more tags related to South Park. But it didn’t focus to title and content. They need more improvements about this tagging issue.
Related posts:
Hi,
Zemanta’s trying hard to automatically suggest tags, and we try to err on the side of abundance of tags. The logics is very simple, you are just clicking, no time is lost for you if we propose a bad tag or two.
However if we wouldn’t propose the good tags, time would be lost because you would need to type them.
But yeah, computers are still not near the human understanding of text. But they are trying
We also have an API that you can leverage to create your own solutions!
bye
Andraz Tori, Zemanta
Then use the API by sending it through google translate first.
Albeit that sounds a bit fishy, we know of people doing it! They report good results (good as in – they can reduce need for typing by having pre-recommendations, naturally there’s more noise)!
bye
Andraz Tori, Zemanta
Pingback: P2P Researchers Fear BitTorrent Meltdown | TorrentFreak | Wyrdpress
Hi,
Thanks for considering and replying my personal thoughts. I really like companies when they pay attention to individuals.
I’ve been interested in and making academic studies about semantic web and I know how hard your business is.
I try to keep my blog in English but my native language is Turkish. Your API may useful for my native language because Zemanta doesn’t work very well in Turkish.
Thanks,
That is neat. Google translate added Turkish language support recently. That would be a nice test for Google translate.