Comments on: MyTV Genie http://blog.stevex.net/2008/01/mytv-genie/ Software development and other notes. Fri, 03 Feb 2012 14:03:53 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Michael Gale http://blog.stevex.net/2008/01/mytv-genie/comment-page-1/#comment-270798 Michael Gale Thu, 17 Jan 2008 17:58:05 +0000 http://blog.stevex.net/index.php/2008/01/17/mytv-genie/#comment-270798 When determining what show to recommend, they could key off of a category index (another interesting key would be time spent watching a particular category). Just because you watch Friends, Seasame Street, and Flip this House, it doesn't necessairily mean I'll see Seasame Street in my recommended lineup (unless I watch Thomas the Train). If they didn't try and match category with category, then it would be difficult to imagine anything other than a seemingly random assortment of recommendations. Some really interesting results could show up, depending on how complex they make their viewing behavior model: user feedback, user/movie ratings, type of content being watched (action, adult, comedy, etc.), and even correlations between television stars and your most watched movies or shows. Imagine getting a recommendation for a movie or television show made ten years ago which features Bruce Campbell because I seem to be a big fan. When determining what show to recommend, they could key off of a category index (another interesting key would be time spent watching a particular category). Just because you watch Friends, Seasame Street, and Flip this House, it doesn’t necessairily mean I’ll see Seasame Street in my recommended lineup (unless I watch Thomas the Train). If they didn’t try and match category with category, then it would be difficult to imagine anything other than a seemingly random assortment of recommendations.

Some really interesting results could show up, depending on how complex they make their viewing behavior model: user feedback, user/movie ratings, type of content being watched (action, adult, comedy, etc.), and even correlations between television stars and your most watched movies or shows. Imagine getting a recommendation for a movie or television show made ten years ago which features Bruce Campbell because I seem to be a big fan.

]]>