public class AverageReciprocalHitRankEvaluator extends AbstractRecommenderEvaluator
Xia Ning and George Karypis, SLIM: Sparse Linear Methods for Top-N Recommender Systems, ICDM 2011.
They apply a leave-one-out validation method to evaluate the algorithm performance. In each run, each of the datasets is split into a training set and a testing set by randomly selecting one of the non-zero entries of each user and placing it into the testing set.
|Constructor and Description|
|Modifier and Type||Method and Description|
Evaluate on the test set with the the list of recommended items.
evaluate, getConf, setTopN
public double evaluate(SparseMatrix testMatrix, RecommendedList recommendedList)
Copyright © 2017. All Rights Reserved.