S and LLL as well as for sparsification techniques such as the
LLR sparsification. All that changes is the mechanics of how you do the
multiplications. Weighting of components works the same as well.
What is very different here is that we have a component of cross
recommendation. That is the B'A
On Thu, Feb 21, 2013 at 12:13 PM, Pat Ferrel wrote:
> I am quite interested in trying this but have a few questions.
>
> To use/abuse mahout to do this:
>
> A and B can be thought of as having the same size, in other words they
> must be constructed to have the same dimension definitions (userID
My plan was to NOT use lucene to start with though I see the benefits. This is
because I want to experiment with weighting--doing idf, no weighting, and with
a non-log idf. Also I want to experiment with temporal decay of recomendability
and maybe blend item similarity based results in certain c
On Fri, Feb 22, 2013 at 9:48 AM, Pat Ferrel wrote:
> ... lots of stuff I can't answer right now ...
>
> Then the question is how to blend B'A h_v with B'B h_p?
>
Add the scores. With coherent weighting schemes that should work just
fine.
To do this, get the two recommendation lists and add s
ce as much work as you need to do and it also doesn't let you
adjust weightings separately.
But it is probably the simplest way to get going with cross recommendation.
On Fri, Feb 22, 2013 at 9:48 AM, Pat Ferrel wrote:
> Therefore my discussion was assuming the use of the entire ma
combined item recommendation matrix which is
roughly twice as much work as you need to do and it also doesn't let you
adjust weightings separately.
But it is probably the simplest way to get going with cross recommendation.
On Fri, Feb 22, 2013 at 9:48 AM, Pat Ferrel wrote:
> There
No.
It is uniformly better to have (item+action, user). In fact, I would
prefer to have it the other way around when describing it to match the
matrix row x column convention.
(user, item+action) where action is binary leads to A = [A_1 | A_2] = user
by 2xitem. The alternative of (user+action,
Yes I understand that you need (user, item+action) input for user based recs
returned from recommender.recommend(userID, n).
But can you expect item similarity to work with the same input? I am fuzzy
about how item similarity is calculated in cf/taste.
I was expecting to train one recommender
On Sun, Feb 24, 2013 at 8:39 AM, Pat Ferrel wrote:
> ...
> But can you expect item similarity to work with the same input? I am fuzzy
> about how item similarity is calculated in cf/taste.
>
The cooccurrence oriented item-based recommenders should be fine.
I was expecting to train one recommend
I may not be 100% following the thread, but:
Similarity metrics won't care whether some items are really actions and
some items are items. The math is the same. The problem which you may be
alluding to is the one I mentioned earlier -- there is no connection
between item and item-action in the mod
Hey,
I have implemented a cross recommender based on the approach Ted Dunning
proposed (cannot find the original post, but here is a follow up
http://www.mail-archive.com/user@mahout.apache.org/msg12983.html).
Currently I am struggling with the last step of blending the initial
recommendations.
I am also very interested in the answer to this question.
Just to reiterate, if you use different recommenders, e.g.,
kNN user-based, kNN item-based, ALS, each one produces
recommendations on a different scale. So how do you combine them?
On Fri, May 31, 2013 at 3:07 PM, Dominik Hübner wrote:
>
In my case, I put all the indicators from all different sources in the same
Solr/Lucene index. Recommendations consists of making a single query to
Solr/Lucene with as much data as I have or want to include.
At the point that this query is done, there are no weights on the
indicators ... merely p
Ted,
Thank you very much. This is very insightful.
The log scaling is definitely an intuitive way of building the meta model.
Not much disagreement about the uselessness of predicting ratings.
On Fri, May 31, 2013 at 4:00 PM, Ted Dunning wrote:
> In my case, I put all the indicators from all di
I've got a cross-recommender too. It was originally conceived to do a
multi-action ensemble from Ted's notes. I'm now gathering a new data set and
building the meta-model learner.
Even with the same scale you need to learn the weighting factors. Consider a
simple ensemble case:
R_p is the matr
Thanks for those detailed responses!
So I assume that the problem of scaling the initial recommendations can be
implicitly solved by learning the weights for the linear combination. The log
rank Ted mentioned seems to be useful as well!
What metric do you use to evaluate your recommendations? I
I use MAP (mean average precision)
http://en.wikipedia.org/wiki/Information_retrieval#Mean_average_precision .
Either with random held out data or more usually a time-based split. The oldest
90% for train the the most recent 10% for test.
We looked at the power of purchases and views for predic
Err, I think there is a mistake below.
You want to do [B'A]H_v, where H_v = A', the user's history of views as column
vectors. At least that is what my code was doing.
On another subject the idea of truncating the user history vectors came up in
another thread. In some research we did using the
I am using [B'A]H_v in my code as well.
Furthermore, the basics of my implementation are done so far (a more "general
purpose" reimplementation of the item-based recommender) and I soon will move
on to evaluate the approach with my dataset. On the one hand I'm quite lucky
since each of the inte
You don't want to decay the training values, only the query values. The
training values indicate user taste similarity and that decays very slowly if
at all. The truncation I was talking about is in the query vectors. And even
with that I'd measure its effect.
If you do this with [B'B]H_p you'
20 matches
Mail list logo