[ https://issues.apache.org/jira/browse/MAHOUT-906?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13169354#comment-13169354 ]
Anatoliy Kats commented on MAHOUT-906: -------------------------------------- You're right, if you have preference values, picking relevant items by recency does not work. But, if you want to test for temporal dynamics, here is a test that works for boolean and rated data: 1. Let days 1...(i-1) be your training data 2. Let day i be your test data 3. Calculate an evaluation metric of the test on a recommender created by the training data. For rated data, you use the difference metric and that's already implemented. For boolean data, you use the IR metric that I proposed earlier: For each user, generate as many recommendations as there are in the test set. Compute IR statistics based on the intersection of actual preferences, and generated recommendations. The easiest way to do that seems to be factoring out the data splitting from the AbstractDifferenceRecommenderEvaluator, which should be renamed, if we go ahead, to reflect its new role. > Allow collaborative filtering evaluators to use custom logic in splitting > data set > ---------------------------------------------------------------------------------- > > Key: MAHOUT-906 > URL: https://issues.apache.org/jira/browse/MAHOUT-906 > Project: Mahout > Issue Type: Improvement > Components: Collaborative Filtering > Affects Versions: 0.5 > Reporter: Anatoliy Kats > Priority: Minor > Labels: features > Original Estimate: 48h > Remaining Estimate: 48h > > I want to start a discussion about factoring out the logic used in splitting > the data set into training and testing. Here is how things stand: There are > two independent evaluator based classes: > AbstractDifferenceRecommenderEvaluator, splits all the preferences randomly > into a training and testing set. GenericRecommenderIRStatsEvaluator takes > one user at a time, removes their top AT preferences, and counts how many of > them the system recommends back. > I have two use cases that both deal with temporal dynamics. In one case, > there may be expired items that can be used for building a training model, > but not a test model. In the other, I may want to simulate the behavior of a > real system by building a preference matrix on days 1-k, and testing on the > ratings the user generated on the day k+1. In this case, it's not items, but > preferences(user, item, rating triplets) which may belong only to the > training set. Before we discuss appropriate design, are there any other use > cases we need to keep in mind? -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira