It looks reasonable to me; what works best for your data will only be revealed with some testing of different algorithms. For example try LogLikelihoodSimilarity.
On Fri, Nov 12, 2010 at 3:41 PM, bejoy ks <[email protected]> wrote: > > Thanks a lot Owen. One more small favor,hope its fine for you. > I'd like to get your suggestion on my implementation of my Item > based recommendation over the same data set.(data set with no > preference value). > > The sample code i have used is given below, it works fine with > recommendations being produced. Could you please look into it and just > verify whether the use of ItemSimilarity and ItemBasedRecommender are the > best suited ones for Boolean data sets. > > > > //item similarity recommendation based on User Id > > FileDataModel dataModel = new FileDataModel(new File(recsFile)); > > ItemSimilarity itemSimilarity = new > TanimotoCoefficientSimilarity(dataModel); > > ItemBasedRecommender recommender =new > GenericItemBasedRecommender(dataModel, itemSimilarity); > > List<RecommendedItem> recommendations > =recommender.recommend(userId, noOfRecommendations); > > System.out.println(recommendations); > > > > //item similarity recommendation based on Item Id > > recommendations= recommender.mostSimilarItems(itemId, > noOfRecommendations); > > System.out.println(recommendations); > > > > Please let me know you suggestions and comments on the code snippet. Is > this the right way to so Item based recommendations on Boolean data > sets? > > Thanks and Regards > Bejoy.K.S > > > > > > Date: Fri, 12 Nov 2010 15:34:03 +0000 > > Subject: Re: Mahout - Help needed - files with no preferences and > integarting mahout with Hadoop > > From: [email protected] > > To: [email protected] > > > > http://manning.com/owen > > > > On Fri, Nov 12, 2010 at 3:16 PM, bejoy ks <[email protected]> wrote: > > > > > > > > Ok that'd be great Owen , if you could point me to the book 'Mahout in > > > Action' . I'm bit interested to know more on the possibilities > available > > > with mahout and also the right usage of similarities, recommenders > >
