spark-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sean Owen <so...@cloudera.com>
Subject Re: MLib: How to set preferences for ALS implicit feedback in Collaborative Filtering?
Date Fri, 16 Jan 2015 15:08:32 GMT
On Fri, Jan 16, 2015 at 9:58 AM, Zork Sail <zorksail@gmail.com> wrote:
> And then train ALSL:
>
>      val model = ALS.trainImplicit(ratings, rank, numIter)
>
> I get RMSE 0.9, which is a big error in case of preferences taking 0 or 1
> value:

This is likely the problem. RMSE is not an appropriate evaluation
metric when you have trained a model on implicit data. The
factorization is not minimizing the same squared error loss that RMSE
evaluates. Use metrics like AUC instead, for example.

Rating value can be 1 if you have no information at all about the
interaction other than that it exists. It should be thought of as a
weight. "10" means it's 10 times more important to predict an
interaction than one with weight "1".

---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@spark.apache.org
For additional commands, e-mail: user-help@spark.apache.org


Mime
View raw message