I guess you think about of having an item similarity algorithms that depends on item features. You only need to implement a new ItemSimilarity as Sean already said. This way, you could make content-based recommendations with non-distrbute Taste. Also, make sure you really want to do this with Taste, as you can also use Mahout clustering algorithms for this.
On 3 September 2011 10:04, Sean Owen <sro...@gmail.com> wrote: > If you are referring to the non-distributed similarity function, it is > quite > easy: implement UserSimilarity or ItemSimilarity and use that as your > similarity function. How you implement is up to you. If you say more about > what you need to do, maybe people can suggest the right logic. > > On Sat, Sep 3, 2011 at 2:26 AM, Walter Chang <weidezhang2...@gmail.com > >wrote: > > > Hi, > > > > If an item has both user rating and item specific contents(category, item > > description content) for example for product recommendation, how can i > > customize the similarity function ? As far as I understand, the current > > mahout similarity function is based on user rating only. Any one had > > experience writing a custom item based similarity function ? > > > > Thanks a lot, > > > > Weide > > > -- -- Marko Ćirić ciric.ma...@gmail.com
