Package moa.tasks

Class EvaluateOnlineRecommender

  • All Implemented Interfaces:
    Configurable, Serializable, MOAObject, OptionHandler, Task

    public class EvaluateOnlineRecommender
    extends AuxiliarMainTask
    Test for evaluating a recommender by training and periodically testing on samples from a rating dataset. When finished, it will show the learning curve of the recommender rating predictor.

    Parameters:

    • d: dataset - the dataset to be used to train/test the rating predictor.
    • f: sample frequency - the frequency in which a rating from the dataset will be used to test the model
    Version:
    $Revision: 7 $
    Author:
    Alex Catarineu (a.catarineu@gmail.com)
    See Also:
    Serialized Form
    • Field Detail

      • ratingPredictorOption

        public ClassOption ratingPredictorOption
      • sampleFrequencyOption

        public IntOption sampleFrequencyOption
    • Constructor Detail

      • EvaluateOnlineRecommender

        public EvaluateOnlineRecommender()
    • Method Detail

      • getTaskResultType

        public Class<?> getTaskResultType()
        Description copied from interface: Task
        Gets the result type of this task. Tasks can return LearningCurve, LearningEvaluation, Classifier, String, Instances..
        Returns:
        a class object of the result of this task
      • doMainTask

        public Object doMainTask​(TaskMonitor monitor,
                                 ObjectRepository repository)
        Description copied from class: MainTask
        This method performs this task. AbstractTask implements doTask, that uses doTaskImpl. MainTask implements doTaskImpl using doMainTask so its extensions only need to implement doMainTask.
        Specified by:
        doMainTask in class MainTask
        Parameters:
        monitor - the TaskMonitor to use
        repository - the ObjectRepository to use
        Returns:
        an object with the result of this task