Concept analysis debuts on Noovo
Posted 1 year ago by Andrej, Viewed 1152 times
Very exciting achievement. We have just deployed our integration with dbpedia, that will introduce item-based aspect to our recommendations. In this way, we might be able to filter out interests that we don't share with our friends and influencers and improve discovery of inherently interesting content.
In basic words, that means that you will get the most relevant recommendation based on topics that we recognize from your individual and social user behaviour (content consumption).
So for example: Kaixin001: China’s Apple of Social Networks. Noovo recognizes the following concepts in this story: world wide web, computing, business, industries, asia, social sciences, sociology, social systems, social networks.
This is very cool! I didn't know about dbpedia at all and I was very impressed once I went to their website. The combination with the recommendation engine of Noovo definitely offers great possibilities. I really believe that "itemized" information that is also described by attributes will help computers become smart and interact in a natural way with humans.
John, 1 year ago, Reply
Sounds interesting. Can you tell us more? Sth like faviki ?
Matjaž, 1 year ago, Reply
We analyze stories for "concepts" (like automated tagging), then query dbpedia where we climb the ontology to discern more generalized categories. The interesting part is that we can feed this stuff to our recommendation engine in the same way as everything else (influencing users, feeds, groups as discerned from your various gestures).
When clicking on light bulbs, you will start seeing these concepts in explanations for some recommendations. Granted, the current implementation is still somewhat rough, so you might get nonsense now and then, but this is something that can be improved. It's a very important first step and we are super excited.
Matej, 1 year ago, Reply
Looks good. Now all we need is more users.
Matjaž, 1 year ago, Reply
Jeah! Good job! :)
ruph, 1 year ago, Reply
Very exciting!
I was not aware that opencyc, Freebase or DBpedia were this accessible.
Neil, 1 year ago, Reply