Package com.google.common.math

Class PairedStats

  • All Implemented Interfaces:
    java.io.Serializable
    public final class PairedStats
extends java.lang.Object
implements java.io.Serializable
    An immutable value object capturing some basic statistics about a collection of paired double values (e.g. points on a plane). Build instances with PairedStatsAccumulator.snapshot().
    Since:
    20.0
    See Also:
    Serialized Form

    • Field Detail

      • xStats

        private final Stats xStats

      • yStats

        private final Stats yStats

      • sumOfProductsOfDeltas

        private final double sumOfProductsOfDeltas

      • BYTES

        private static final int BYTES
        The size of byte array representation in bytes.
        See Also:
        Constant Field Values

    • Constructor Detail

      • PairedStats

        PairedStats​(Stats xStats,
            Stats yStats,
            double sumOfProductsOfDeltas)
        Internal constructor. Users should use PairedStatsAccumulator.snapshot().

        To ensure that the created instance obeys its contract, the parameters should satisfy the following constraints. This is the callers responsibility and is not enforced here.

        • Both xStats and yStats must have the same count.
        • If that count is 1, sumOfProductsOfDeltas must be exactly 0.0.
        • If that count is more than 1, sumOfProductsOfDeltas must be finite.

    • Method Detail

      • count

        public long count()
        Returns the number of pairs in the dataset.

      • xStats

        public Stats xStats()
        Returns the statistics on the x values alone.

      • yStats

        public Stats yStats()
        Returns the statistics on the y values alone.

      • populationCovariance

        public double populationCovariance()
        Returns the population covariance of the values. The count must be non-zero.

        This is guaranteed to return zero if the dataset contains a single pair of finite values. It is not guaranteed to return zero when the dataset consists of the same pair of values multiple times, due to numerical errors.

        Non-finite values

        If the dataset contains any non-finite values (Double.POSITIVE_INFINITY, Double.NEGATIVE_INFINITY, or Double.NaN) then the result is Double.NaN.

        Throws:
        java.lang.IllegalStateException - if the dataset is empty

      • sampleCovariance

        public double sampleCovariance()
        Returns the sample covariance of the values. The count must be greater than one.

        This is not guaranteed to return zero when the dataset consists of the same pair of values multiple times, due to numerical errors.

        Non-finite values

        If the dataset contains any non-finite values (Double.POSITIVE_INFINITY, Double.NEGATIVE_INFINITY, or Double.NaN) then the result is Double.NaN.

        Throws:
        java.lang.IllegalStateException - if the dataset is empty or contains a single pair of values

      • pearsonsCorrelationCoefficient

        public double pearsonsCorrelationCoefficient()
        Returns the Pearson's or product-moment correlation coefficient of the values. The count must greater than one, and the x and y values must both have non-zero population variance (i.e. xStats().populationVariance() > 0.0 && yStats().populationVariance() > 0.0). The result is not guaranteed to be exactly +/-1 even when the data are perfectly (anti-)correlated, due to numerical errors. However, it is guaranteed to be in the inclusive range [-1, +1].

        Non-finite values

        If the dataset contains any non-finite values (Double.POSITIVE_INFINITY, Double.NEGATIVE_INFINITY, or Double.NaN) then the result is Double.NaN.

        Throws:
        java.lang.IllegalStateException - if the dataset is empty or contains a single pair of values, or either the x and y dataset has zero population variance

      • leastSquaresFit

        public LinearTransformation leastSquaresFit()
        Returns a linear transformation giving the best fit to the data according to Ordinary Least Squares linear regression of y as a function of x. The count must be greater than one, and either the x or y data must have a non-zero population variance (i.e. xStats().populationVariance() > 0.0 || yStats().populationVariance() > 0.0). The result is guaranteed to be horizontal if there is variance in the x data but not the y data, and vertical if there is variance in the y data but not the x data.

        This fit minimizes the root-mean-square error in y as a function of x. This error is defined as the square root of the mean of the squares of the differences between the actual y values of the data and the values predicted by the fit for the x values (i.e. it is the square root of the mean of the squares of the vertical distances between the data points and the best fit line). For this fit, this error is a fraction sqrt(1 - R*R) of the population standard deviation of y, where R is the Pearson's correlation coefficient (as given by pearsonsCorrelationCoefficient()).

        The corresponding root-mean-square error in x as a function of y is a fraction sqrt(1/(R*R) - 1) of the population standard deviation of x. This fit does not normally minimize that error: to do that, you should swap the roles of x and y.

        Non-finite values

        If the dataset contains any non-finite values (Double.POSITIVE_INFINITY, Double.NEGATIVE_INFINITY, or Double.NaN) then the result is LinearTransformation.forNaN().

        Throws:
        java.lang.IllegalStateException - if the dataset is empty or contains a single pair of values, or both the x and y dataset must have zero population variance

      • equals

        public boolean equals​(java.lang.Object obj)

        Note: This tests exact equality of the calculated statistics, including the floating point values. Two instances are guaranteed to be considered equal if one is copied from the other using second = new PairedStatsAccumulator().addAll(first).snapshot(), if both were obtained by calling snapshot() on the same PairedStatsAccumulator without adding any values in between the two calls, or if one is obtained from the other after round-tripping through java serialization. However, floating point rounding errors mean that it may be false for some instances where the statistics are mathematically equal, including instances constructed from the same values in a different order... or (in the general case) even in the same order. (It is guaranteed to return true for instances constructed from the same values in the same order if strictfp is in effect, or if the system architecture guarantees strictfp-like semantics.)

        Overrides:
        equals in class java.lang.Object

      • hashCode

        public int hashCode()

        Note: This hash code is consistent with exact equality of the calculated statistics, including the floating point values. See the note on equals(java.lang.Object) for details.

        Overrides:
        hashCode in class java.lang.Object

      • toString

        public java.lang.String toString()
        Overrides:
        toString in class java.lang.Object

      • sumOfProductsOfDeltas

        double sumOfProductsOfDeltas()

      • ensurePositive

        private static double ensurePositive​(double value)

      • ensureInUnitRange

        private static double ensureInUnitRange​(double value)

      • toByteArray

        public byte[] toByteArray()
        Gets a byte array representation of this instance.

        Note: No guarantees are made regarding stability of the representation between versions.

      • fromByteArray

        public static PairedStats fromByteArray​(byte[] byteArray)
        Creates a PairedStats instance from the given byte representation which was obtained by toByteArray().

        Note: No guarantees are made regarding stability of the representation between versions.