You just use the same value for all ratings, like '1.0' v.s. '0.0' for no rating found.
On Tue, Feb 21, 2012 at 1:25 PM, Manuel Blechschmidt <[email protected]> wrote: > Hi Simon, > yes, this is possible. The following demo project is doing that: > > https://github.com/ManuelB/facebook-recommender-demo > > If you replace the Demo file with your data file you are ready to go. > https://github.com/ManuelB/facebook-recommender-demo/blob/master/src/main/resources/DemoFriendsLikes.csv > (Actually only the first to entries are needed) > > Here is a well documented EJB which is doing the complete recommender part: > https://github.com/ManuelB/facebook-recommender-demo/blob/master/src/main/java/de/apaxo/bedcon/FacebookRecommender.java > > Hope that helps > Manuel > > On 21.02.2012, at 19:47, Simon Franz wrote: > >> Hi there, >> I'm very new to Mahout and recommender systems. I have made my first >> steps with mahout in java and mahout is really nice and quite "simple" >> to use. >> In my scenario I want to use Mahout as recommender for music. I have >> many tracks and users and a set of what each user has listen at a >> particular timestamp. Does Mahout support a recommender who can >> recommender for a datafile without ratings? Something like >> <userid>,<trackid> ? >> >> Thanks in advance. >> Best Regards >> Simon > > -- > Manuel Blechschmidt > Dortustr. 57 > 14467 Potsdam > Mobil: 0173/6322621 > Twitter: http://twitter.com/Manuel_B > -- Lance Norskog [email protected]
