Package adams.gui.tools.wekainvestigator.tab.classifytab

Class PredictionHelper

  • public class PredictionHelper
extends Object
    Helper class for dealing with predictions from result items.
    Author:
    FracPete (fracpete at waikato dot ac dot nz)

    • Constructor Detail

      • PredictionHelper

        public PredictionHelper()

    • Method Detail

      • toSubset

        public static int[] toSubset​(int[] originalIndices)
        Treats the indices as a subset, generating new indices while maintaining order.
        Parameters:
        originalIndices - the indices to process
        Returns:
        the indices for the subset

      • toSubset

        public static SpreadSheet toSubset​(int[] originalIndices,
                                   SpreadSheet additionalAttributes)
        Returns the subset of the additional attributes as per indices.
        Parameters:
        originalIndices - the indices of the subset
        additionalAttributes - the full additional attributes
        Returns:
        the subset of the additional attributes

      • toSpreadSheet

        public static SpreadSheet toSpreadSheet​(LoggingSupporter logger,
                                        MessageCollection errors,
                                        ResultItem item,
                                        boolean addAdditionalAttributes,
                                        boolean showError)
        Turns the result item into a spreadsheet with the predictions.
        Parameters:
        logger - the object used for logging, can be null
        errors - for collecting errors
        item - the result item to use
        addAdditionalAttributes - whether to add additional attributes
        showError - whether to add the error in a separate column
        Returns:
        the generated spreadsheet

      • toSpreadSheet

        public static SpreadSheet toSpreadSheet​(LoggingSupporter logger,
                                        MessageCollection errors,
                                        weka.classifiers.Evaluation eval,
                                        int[] originalIndices,
                                        SpreadSheet additionalAttributes,
                                        boolean showError)
        Turns the result item into a spreadsheet with the predictions.
        Parameters:
        logger - the object used for logging, can be null
        errors - for collecting errors
        eval - the evaluation to use
        originalIndices - the original indices to use, can be null
        additionalAttributes - the additional attributes to add, can be null
        showError - whether to add the error in a separate column
        Returns:
        the generated spreadsheet

      • toSpreadSheet

        public static SpreadSheet toSpreadSheet​(LoggingSupporter logger,
                                        MessageCollection errors,
                                        ResultItem item,
                                        boolean addAdditionalAttributes,
                                        boolean addLabelIndex,
                                        boolean showDistribution,
                                        boolean showProbability,
                                        boolean showError,
                                        boolean showWeight)
        Turns the result item into a spreadsheet with the predictions.
        Parameters:
        logger - the object used for logging, can be null
        errors - for collecting errors
        item - the result item to use
        addAdditionalAttributes - whether to add additional attributes
        addLabelIndex - whether to add the label index in a separate column
        showDistribution - whether to add the distribution in a separate column
        showProbability - whether to add the probability in a separate column
        showError - whether to add the error in a separate column
        showWeight - whether to add the weight in a separate column
        Returns:
        the generated spreadsheet, null if failed

      • toSpreadSheet

        public static SpreadSheet toSpreadSheet​(LoggingSupporter logger,
                                        MessageCollection errors,
                                        weka.classifiers.Evaluation eval,
                                        int[] originalIndices,
                                        SpreadSheet additionalAttributes,
                                        boolean addLabelIndex,
                                        boolean showDistribution,
                                        boolean showProbability,
                                        boolean showError,
                                        boolean showWeight)
        Turns the result item into a spreadsheet with the predictions.
        Parameters:
        logger - the object used for logging, can be null
        errors - for collecting errors
        eval - the evaluation to use
        originalIndices - the original indices to use, can be null
        additionalAttributes - the additional attributes to use, can be null
        addLabelIndex - whether to add the label index in a separate column
        showDistribution - whether to add the distribution in a separate column
        showProbability - whether to add the probability in a separate column
        showError - whether to add the error in a separate column
        showWeight - whether to add the weight in a separate column
        Returns:
        the generated spreadsheet, null if failed

      • toSpreadSheet

        public static SpreadSheet toSpreadSheet​(LoggingSupporter logger,
                                        MessageCollection errors,
                                        weka.classifiers.Evaluation eval,
                                        int[] originalIndices,
                                        SpreadSheet additionalAttributes,
                                        boolean addLabelIndex,
                                        boolean showDistribution,
                                        boolean showProbability,
                                        boolean showError,
                                        boolean showWeight,
                                        boolean showAbsError,
                                        boolean showRelError)
        Turns the result item into a spreadsheet with the predictions.
        Parameters:
        logger - the object used for logging, can be null
        errors - for collecting errors
        eval - the evaluation to use
        originalIndices - the original indices to use, can be null
        additionalAttributes - the additional attributes to use, can be null
        addLabelIndex - whether to add the label index in a separate column
        showDistribution - whether to add the distribution in a separate column
        showProbability - whether to add the probability in a separate column
        showError - whether to add the error in a separate column
        showWeight - whether to add the weight in a separate column
        showAbsError - whether to show absolute errors
        showRelError - whether to show the relative error
        Returns:
        the generated spreadsheet, null if failed