Package com.tibco.patterns.learn.recpair

Class RecPairExperiment

  • public final class RecPairExperiment
extends RLinkExperiment<RecPairExample>
    Manages a training dataset and a validation dataset. The datasets consist of record pairs.

    • Constructor Summary

      Constructors 
      Constructor Description
      RecPairExperiment​(double vldProportion, java.lang.String recPairCsvFileName, RecPairScorer scorer)
      Loads a single list of pairs from CSV, then divides them randomly to training and validation datasets.
      RecPairExperiment​(int nFeatures, java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlTrainDataset, java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlVldDataset)
      Creates RecPairExperiment with training examples imported from the two given XML datasets.

    • Method Summary

      All Methods Instance Methods Concrete Methods 
      Modifier and Type Method Description
      void copyModelResultsToXml​(java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlTrainDataset, java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlVldDataset, com.tibco.patterns.learn.jaxb.model.ObjectFactory factory)
      Copies RLink predictions for all examples to XML datasets.

      • Methods inherited from class java.lang.Object

        clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    • Constructor Detail

      • RecPairExperiment

        public RecPairExperiment​(double vldProportion,
                         java.lang.String recPairCsvFileName,
                         RecPairScorer scorer)
                  throws com.netrics.likeit.NetricsFileFormatException,
                         java.io.IOException,
                         com.netrics.likeit.NetricsException
        Loads a single list of pairs from CSV, then divides them randomly to training and validation datasets. Calculates symmetric feature scores for each pair. Each line in CSV has format "Key1,Key2,label". It cannot contain spaces before/after a comma. The label is 0 or 1. Pairs with other label values are skipped.
        Parameters:
        vldProportion - - probability that an example is assigned to validation dataset.
        recPairCsvFileName - - CSV file name.
        scorer - - object used to calculate feature scores for each pair.
        Throws:
        java.lang.IllegalArgumentException - if vldProportion is not in interval [0.0, 1.0), or if number of query field values does not match the number of all fields in feature query.
        java.lang.IllegalStateException - if server table was not provided when creating scorer.
        com.netrics.likeit.NetricsFileFormatException - - CSV file format errors
        com.netrics.likeit.NetricsException - if an error occurred on server
        java.io.IOException - if number of entries per line is wrong; or error communicating with server

      • RecPairExperiment

        public RecPairExperiment​(int nFeatures,
                         java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlTrainDataset,
                         java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlVldDataset)
        Creates RecPairExperiment with training examples imported from the two given XML datasets. To be used from UI.
        Parameters:
        nFeatures - - number of model features in each example of this experiment
        xmlTrainDataset - - training dataset in XML format.
        xmlVldDataset - - validation dataset in XML format.

    • Method Detail

      • copyModelResultsToXml

        public void copyModelResultsToXml​(java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlTrainDataset,
                                  java.util.List<com.tibco.patterns.learn.jaxb.model.RecPairType> xmlVldDataset,
                                  com.tibco.patterns.learn.jaxb.model.ObjectFactory factory)
        Copies RLink predictions for all examples to XML datasets. To be used from UI. The given XML datasets must not be modified since creating this experiment.
        Parameters:
        xmlTrainDataset - - training dataset in XML format.
        xmlVldDataset - - validation dataset in XML format.
        factory - - used to create PredictionType for RecPairs