http://socialfabric.co.uk/2009/01/recommendation-engines-a-bit-of-give-and-tak/ digital blog Fri, 09 Oct 2009 15:12:01 +0000 hourly 1 http://wordpress.org/?v=3.0 http://socialfabric.co.uk/2009/01/recommendation-engines-a-bit-of-give-and-tak/comment-page-1/#comment-6 neil Sun, 11 Jan 2009 14:14:23 +0000 http://blog.socialfabric.co.uk/?p=31#comment-6 It is that human perception element that is the stumbling block - currently. The approach of using the group to improve recommendations for the individual is compelling. Bit like Google's SearchWiki and <a href="http://www.flickr.com/commons" rel="nofollow">Flickr's The Commons</a> amongst others. I suppose the purists would say that for the Netflix challenge it is about providing an algorithm that is able to work independent of crowdsourcing style techniques. It is that human perception element that is the stumbling block - currently. The approach of using the group to improve recommendations for the individual is compelling. Bit like Google’s SearchWiki and Flickr’s The Commons amongst others. I suppose the purists would say that for the Netflix challenge it is about providing an algorithm that is able to work independent of crowdsourcing style techniques.

]]> http://socialfabric.co.uk/2009/01/recommendation-engines-a-bit-of-give-and-tak/comment-page-1/#comment-5 Phoebe Sun, 11 Jan 2009 09:08:37 +0000 http://blog.socialfabric.co.uk/?p=31#comment-5 One of the most interesting quotes from the Netflix article is "Hastings is even considering hiring cinephiles to watch all 100,000 movies in the Netflix library and write up, by hand, pages of adjectives describing each movie." Perhaps this extra data is needed to significantly improve Netflix recommendations? That's our approach at Jinni (http://www.jinni.com) - our search and recommendations are based on the Movie Genome. One of the most interesting quotes from the Netflix article is “Hastings is even considering hiring cinephiles to watch all 100,000 movies in the Netflix library and write up, by hand, pages of adjectives describing each movie.” Perhaps this extra data is needed to significantly improve Netflix recommendations? That’s our approach at Jinni (http://www.jinni.com) - our search and recommendations are based on the Movie Genome.

]]>