Skip navigation links
org.deidentifier.arx.metric.v2

Class __MetricV2

    • Constructor Detail

      • __MetricV2

        public __MetricV2()
    • Method Detail

      • createAECSMetric

        public static Metric<ILSingleDimensional> createAECSMetric()
        Creates a new instance of the AECS metric.
        Returns:
      • createAECSMetric

        public static Metric<ILSingleDimensional> createAECSMetric(double gsFactor)
        Creates a new instance of the AECS metric.
        Parameters:
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createAECSMetric

        public static Metric<ILSingleDimensional> createAECSMetric(int rowCount)
        Creates a new instance of the AECS metric.
        Parameters:
        rowCount -
        Returns:
      • createAmbiguityMetric

        public static Metric<ILSingleDimensional> createAmbiguityMetric()
        Creates an instance of the ambiguity metric.
        Returns:
      • createClassificationMetric

        public static Metric<ILSingleDimensional> createClassificationMetric(double gsFactor)
        Creates an instance of the classification metric.
        Parameters:
        gsFactor -
        Returns:
      • createClassificationMetric

        public static Metric<ILSingleDimensional> createClassificationMetric()
        Creates an instance of the classification metric.
        Returns:
      • createDiscernabilityMetric

        public static Metric<ILSingleDimensional> createDiscernabilityMetric()
        Creates an instance of the discernability metric.
        Returns:
      • createDiscernabilityMetric

        public static Metric<ILSingleDimensional> createDiscernabilityMetric(boolean monotonic)
        Creates an instance of the discernability metric. The monotonic variant is DM*.
        Parameters:
        monotonic - If set to true, the monotonic variant (DM*) will be created
        Returns:
      • createDiscernabilityMetric

        public static Metric<ILSingleDimensional> createDiscernabilityMetric(boolean monotonic,
                                                                             double numTuples)
        Creates an instance of the discernability metric. The monotonic variant is DM*.
        Parameters:
        monotonic - If set to true, the monotonic variant (DM*) will be created
        numTuples - Pre-initialization
        Returns:
      • createEntropyBasedInformationLossMetric

        public static MetricSDNMEntropyBasedInformationLoss createEntropyBasedInformationLossMetric(double gsFactor)
        Creates an instance of the entropy-based information loss metric.
        Parameters:
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createEntropyMetric

        public static Metric<AbstractILMultiDimensional> createEntropyMetric()
        Creates an instance of the non-monotonic non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Returns:
      • createEntropyMetric

        public static Metric<AbstractILMultiDimensional> createEntropyMetric(boolean monotonic)
        Creates an instance of the non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        Returns:
      • createEntropyMetric

        public static Metric<AbstractILMultiDimensional> createEntropyMetric(boolean monotonic,
                                                                             Metric.AggregateFunction function)
        Creates an instance of the non-uniform entropy metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        function - The aggregate function to be used for comparing results
        Returns:
      • createEntropyMetric

        public static Metric<AbstractILMultiDimensional> createEntropyMetric(boolean monotonic,
                                                                             double gsFactor)
        Creates an instance of the non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createEntropyMetric

        public static Metric<AbstractILMultiDimensional> createEntropyMetric(boolean monotonic,
                                                                             double gsFactor,
                                                                             Metric.AggregateFunction function)
        Creates an instance of the non-uniform entropy metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        function - The aggregate function to be used for comparing results
        Returns:
      • createEntropyMetric

        public static Metric<AbstractILMultiDimensional> createEntropyMetric(boolean monotonic,
                                                                             double[][] cache,
                                                                             int[][][] cardinalities,
                                                                             int[][][] hierarchies)
        Creates an instance of the non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        cache -
        cardinalities -
        hierarchies -
        Returns:
      • createEntropyMetric

        public static Metric<AbstractILMultiDimensional> createEntropyMetric(double gsFactor)
        Creates an instance of the non-monotonic non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createHeightMetric

        public static Metric<AbstractILMultiDimensional> createHeightMetric()
        Creates an instance of the height metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Returns:
      • createHeightMetric

        public static Metric<AbstractILMultiDimensional> createHeightMetric(Metric.AggregateFunction function)
        Creates an instance of the height metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        function - The aggregate function to use for comparing results
        Returns:
      • createHeightMetric

        public static Metric<AbstractILMultiDimensional> createHeightMetric(int minHeight,
                                                                            int maxHeight)
        Creates an instance of the height metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        minHeight -
        maxHeight -
        Returns:
      • createILMultiDimensionalArithmeticMean

        public static InformationLoss<?> createILMultiDimensionalArithmeticMean(double value)
        Helper method. Normally, there should be no need to call this
        Parameters:
        value -
        Returns:
      • createILMultiDimensionalSum

        public static InformationLoss<?> createILMultiDimensionalSum(double value)
        Helper method. Normally, there should be no need to call this
        Parameters:
        value -
        Returns:
      • createILSingleDimensional

        public static InformationLoss<?> createILSingleDimensional(double value)
        Helper method. Normally, there should be no need to call this
        Parameters:
        value -
        Returns:
      • createKLDivergenceMetric

        public static Metric<ILSingleDimensional> createKLDivergenceMetric()
        Creates an instance of the KL Divergence metric.
        Returns:
      • createLossMetric

        public static Metric<AbstractILMultiDimensional> createLossMetric()
        Creates an instance of the loss metric which treats generalization and suppression equally. The default aggregate function, which is the geometric mean, will be used. This metric will respect attribute weights defined in the configuration.
        Returns:
      • createLossMetric

        public static Metric<AbstractILMultiDimensional> createLossMetric(Metric.AggregateFunction function)
        Creates an instance of the loss metric which treats generalization and suppression equally. This metric will respect attribute weights defined in the configuration.
        Parameters:
        function - The aggregate function to use for comparing results
        Returns:
      • createLossMetric

        public static Metric<AbstractILMultiDimensional> createLossMetric(double gsFactor)
        Creates an instance of the loss metric with factors for weighting generalization and suppression. The default aggregate function, which is the geometric mean, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createLossMetric

        public static Metric<AbstractILMultiDimensional> createLossMetric(double gsFactor,
                                                                          Metric.AggregateFunction function)
        Creates an instance of the loss metric with factors for weighting generalization and suppression. This metric will respect attribute weights defined in the configuration.
        Parameters:
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        function - The aggregate function to use for comparing results
        Returns:
      • createNormalizedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createNormalizedEntropyMetric()
        Creates an instance of the normalized entropy metric. The default aggregate function, which is the sum function, will be used. This metric will respect attribute weights defined in the configuration.
        Returns:
      • createNormalizedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createNormalizedEntropyMetric(Metric.AggregateFunction function)
        Creates an instance of the normalized entropy metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        function - The aggregate function to use for comparing results
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric()
        Creates an instance of the non-monotonic precision metric. The default aggregate function, which is the arithmetic mean, will be used. This metric will respect attribute weights defined in the configuration.
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(Metric.AggregateFunction function)
        Creates an instance of the non-monotonic precision metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        function - The aggregate function to use for comparing results
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(boolean monotonic)
        Creates an instance of the precision metric. The default aggregate function, which is the arithmetic mean, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(boolean monotonic,
                                                                               Metric.AggregateFunction function)
        Creates an instance of the precision metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        function -
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(boolean monotonic,
                                                                               double gsFactor)
        Creates an instance of the precision metric. The default aggregate function, which is the arithmetic mean, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(boolean monotonic,
                                                                               double gsFactor,
                                                                               Metric.AggregateFunction function)
        Creates an instance of the precision metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        function -
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(boolean monotonic,
                                                                               int[] heights,
                                                                               double cells)
        Creates an instance of the precision metric. The default aggregate function, which is the arithmetic mean, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        monotonic - If set to true, the monotonic variant of the metric will be created
        heights -
        cells -
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(double gsFactor)
        Creates an instance of the non-monotonic precision metric. The default aggregate function, which is the arithmetic mean, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createPrecisionMetric

        public static Metric<AbstractILMultiDimensional> createPrecisionMetric(double gsFactor,
                                                                               Metric.AggregateFunction function)
        Creates an instance of the non-monotonic precision metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        function - The aggregate function to use for comparing results
        Returns:
      • createPrecomputedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedEntropyMetric(double threshold)
        Creates a potentially precomputed instance of the non-monotonic non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        Returns:
      • createPrecomputedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedEntropyMetric(double threshold,
                                                                                        boolean monotonic)
        Creates a potentially precomputed instance of the non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        monotonic - If set to true, the monotonic variant of the metric will be created
        Returns:
      • createPrecomputedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedEntropyMetric(double threshold,
                                                                                        boolean monotonic,
                                                                                        Metric.AggregateFunction function)
        Creates a potentially precomputed instance of the non-uniform entropy metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        monotonic - If set to true, the monotonic variant of the metric will be created
        function - The aggregate function to be used for comparing results
        Returns:
      • createPrecomputedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedEntropyMetric(double threshold,
                                                                                        boolean monotonic,
                                                                                        double gsFactor)
        Creates a potentially precomputed instance of the non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        monotonic - If set to true, the monotonic variant of the metric will be created
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createPrecomputedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedEntropyMetric(double threshold,
                                                                                        boolean monotonic,
                                                                                        double gsFactor,
                                                                                        Metric.AggregateFunction function)
        Creates a potentially precomputed instance of the non-uniform entropy metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        monotonic - If set to true, the monotonic variant of the metric will be created
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        function - The aggregate function to be used for comparing results
        Returns:
      • createPrecomputedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedEntropyMetric(double threshold,
                                                                                        double gsFactor)
        Creates a potentially precomputed instance of the non-monotonic non-uniform entropy metric. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createPrecomputedLossMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedLossMetric(double threshold)
        Creates a potentially precomputed instance of the loss metric which treats generalization and suppression equally. The default aggregate function, which is the geometric mean, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        Returns:
      • createPrecomputedLossMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedLossMetric(double threshold,
                                                                                     Metric.AggregateFunction function)
        Creates a potentially precomputed instance of the loss metric which treats generalization and suppression equally. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        function - The aggregate function to use for comparing results
        Returns:
      • createPrecomputedLossMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedLossMetric(double threshold,
                                                                                     double gsFactor)
        Creates a potentially precomputed instance of the loss metric with factors for weighting generalization and suppression. The default aggregate function, which is the gemetric mean, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createPrecomputedLossMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedLossMetric(double threshold,
                                                                                     double gsFactor,
                                                                                     Metric.AggregateFunction function)
        Creates a potentially precomputed instance of the loss metric with factors for weighting generalization and suppression. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        function - The aggregate function to use for comparing results
        Returns:
      • createPrecomputedNormalizedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedNormalizedEntropyMetric(double threshold)
        Creates a potentially precomputed instance of the normalized entropy metric. The default aggregate function, which is the sum function, will be used. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        Returns:
      • createPrecomputedNormalizedEntropyMetric

        public static Metric<AbstractILMultiDimensional> createPrecomputedNormalizedEntropyMetric(double threshold,
                                                                                                  Metric.AggregateFunction function)
        Creates a potentially precomputed instance of the normalized entropy metric. This metric will respect attribute weights defined in the configuration.
        Parameters:
        threshold - The precomputed variant of the metric will be used if #distinctValues / #rows <= threshold for all quasi-identifiers.
        function - The aggregate function to use for comparing results
        Returns:
      • createPublisherBenefitMetric

        public static MetricSDNMPublisherPayout createPublisherBenefitMetric(boolean journalistAttackerModel,
                                                                             double gsFactor)
        Creates an instance of the model for maximizing publisher benefit in the game-theoretic privacy model based on a cost/benefit analysis.
        Parameters:
        journalistAttackerModel - If set to true, the journalist attacker model will be assumed, the prosecutor model will be assumed, otherwise
        gsFactor - A factor [0,1] weighting generalization and suppression. The default value is 0.5, which means that generalization and suppression will be treated equally. A factor of 0 will favor suppression, and a factor of 1 will favor generalization. The values in between can be used for balancing both methods.
        Returns:
      • createStaticMetric

        public static Metric<AbstractILMultiDimensional> createStaticMetric(java.util.Map<java.lang.String,java.util.List<java.lang.Double>> loss)
        Creates an instance of a metric with statically defined information loss. The default aggregate function, which is the sum-function, will be used for comparing results. This metric will respect attribute weights defined in the configuration.
        Parameters:
        loss - User defined information loss per attribute
        Returns:
      • createStaticMetric

        public static Metric<AbstractILMultiDimensional> createStaticMetric(java.util.Map<java.lang.String,java.util.List<java.lang.Double>> loss,
                                                                            Metric.AggregateFunction function)
        Creates an instance of a metric with statically defined information loss. This metric will respect attribute weights defined in the configuration.
        Parameters:
        loss - User defined information loss per attribute
        function - The aggregate function to use for comparing results
        Returns: