Also, back to the issue of the PlusAnonymousUserDataModel breaking the recommender, I tracked the issue down to being a mistake on my part. When I switched to the PlusAnonymousUserDataModel, I added the anonModel to the constructor of my recommender, but accidentally left the original data model in the constructors for my UserSimilarity, and UserNeighborhood.
It didn't throw any exceptions, but since the UserSimilarity and UserNeighborhood had data models that didn't include the anon user, the UserNeighborhood always returned an empty array for the anon user, which lead to the recommender returning 0 recommendations. On Mon, Aug 12, 2013 at 9:47 AM, Matt Molek <[email protected]> wrote: > I'm using a custom PlusAnonUser recommender which is just a > GenericBooleanPrefUserBasedRecommender with a PlusAnonymousUser DataModel > wrapped around a GenericBooleanPrefDataModel, and a > NearestNUserNeighborhood of 25. Similarity measure is LogLikelihood. > > Thanks for the help! > > > On Fri, Aug 9, 2013 at 4:13 PM, Ted Dunning <[email protected]> wrote: > >> On Fri, Aug 9, 2013 at 12:30 PM, Matt Molek <[email protected]> wrote: >> >> > From some local IR precision/recall testing, I've found that user based >> > recommenders do better on my data, so I'd like to stick with user based >> if >> > I can. I know precision/recall measures aren't always that important >> when >> > dealing with recommendation, but in the case I'm using the recommender >> for, >> > I think it's worth maximizing. I'm getting more than double the >> precision >> > out of the user based recommenders. >> > >> >> What kind of user based recommender are you using? >> >> Most competitive user based recommenders can be restated as item-based >> recommenders. Those are much easier to deploy. >> > >
