Package com.oracle.coherence.ai.util

Class Vectors

java.lang.Object
com.oracle.coherence.ai.util.Vectors
public class Vectors extends Object
Utility methods for supporting vectors.

  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    static final float
    EPSILON
    A very small value to use to avoid divide by zero errors.

  • Constructor Summary

    Constructors
    Constructor
    Description
    Vectors()
     

  • Method Summary

    Modifier and Type
    Method
    Description
    static BitVector
    binaryQuant(byte[] vector)
    Quantize an Int8 vector to a bit vector.
    static BitVector
    binaryQuant(float[] vector)
    Quantize a float vector to a bit vector.
    static double
    dotProduct(byte[] v1, byte[] v2)
    Calculate the dot product of two Int8 vectors.
    static double
    dotProduct(float[] v1, float[] v2)
    Calculate the dot product of two float vectors.
    static double
    dotProduct(BitSet v1, BitSet v2)
    Calculate the dot product of two bit vectors.
    static int
    hammingDistance(Vector<BitSet> x, Vector<BitSet> y)
    Calculate the hamming distance between two bit vectors.
    static int
    hammingDistance(BitSet x, BitSet y)
    Calculate the hamming distance between two bit vectors.
    static double
    l2squared(byte[] v1, byte[] v2)
    Calculate the L2 Squared value for two Int8 vectors.
    static double
    l2squared(float[] v1, float[] v2)
    Calculate the L2 Squared value for two float vectors.
    static double
    l2squared(BitSet v1, BitSet v2)
    Calculate the L2 Squared value for two bit vectors.
    static double
    magnitude(byte[] v)
    Calculate the magnitude of an Int8 vector.
    static double
    magnitude(float[] v)
    Calculate the magnitude of a float vector.
    static double
    magnitude(BitSet v)
    Calculate the magnitude of a bit vector.
    static float[]
    normalize(byte[] array)
    Normalize an Int8 vector.
    static float[]
    normalize(float[] array)
    Normalize a float vector.

    Methods inherited from class java.lang.Object

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

  • Field Details

    • EPSILON

      public static final float EPSILON
      A very small value to use to avoid divide by zero errors.
      See Also:

  • Constructor Details

    • Vectors

      public Vectors()

  • Method Details

    • magnitude

      public static double magnitude(BitSet v)
      Calculate the magnitude of a bit vector.
      Parameters:
      v - the bit vector
      Returns:
      the magnitude of the bit vector

    • magnitude

      public static double magnitude(byte[] v)
      Calculate the magnitude of an Int8 vector.
      Parameters:
      v - the Int8 vector
      Returns:
      the magnitude of the Int8 vector

    • magnitude

      public static double magnitude(float[] v)
      Calculate the magnitude of a float vector.
      Parameters:
      v - the float vector
      Returns:
      the magnitude of the float vector

    • dotProduct

      public static double dotProduct(BitSet v1, BitSet v2)
      Calculate the dot product of two bit vectors.
      Parameters:
      v1 - the first bit vector
      v2 - the second bit vector
      Returns:
      the dot product of the bit vectors

    • dotProduct

      public static double dotProduct(byte[] v1, byte[] v2)
      Calculate the dot product of two Int8 vectors.
      Parameters:
      v1 - the first Int8 vector
      v2 - the second Int8 vector
      Returns:
      the dot product of the bit vectors

    • dotProduct

      public static double dotProduct(float[] v1, float[] v2)
      Calculate the dot product of two float vectors.
      Parameters:
      v1 - the first float vector
      v2 - the second float vector
      Returns:
      the magnitude of the float vectors

    • l2squared

      public static double l2squared(BitSet v1, BitSet v2)
      Calculate the L2 Squared value for two bit vectors.
      Parameters:
      v1 - the first bit vector
      v2 - the second bit vector
      Returns:
      the L2 Squared value for the two bit vectors

    • l2squared

      public static double l2squared(byte[] v1, byte[] v2)
      Calculate the L2 Squared value for two Int8 vectors.
      Parameters:
      v1 - the first Int8 vector
      v2 - the second Int8 vector
      Returns:
      the L2 Squared value for the two bit vectors

    • l2squared

      public static double l2squared(float[] v1, float[] v2)
      Calculate the L2 Squared value for two float vectors.
      Parameters:
      v1 - the first float vector
      v2 - the second float vector
      Returns:
      the L2 Squared value for the two bit vectors

    • normalize

      public static float[] normalize(byte[] array)
      Normalize an Int8 vector.
      Parameters:
      array - the Int8 vector to normalize.
      Returns:
      the Int8 vector normalized to a float vector

    • normalize

      public static float[] normalize(float[] array)
      Normalize a float vector.

      Note, the vector is normalized in place, so the values in the array parameter will be updated and the same array returned.

      Parameters:
      array - the float vector to normalize.
      Returns:
      the normalized to a float vector

    • binaryQuant

      public static BitVector binaryQuant(float[] vector)
      Quantize a float vector to a bit vector.
      Parameters:
      vector - the float vector to quantize
      Returns:
      the result of the binary quantization of the float vector

    • binaryQuant

      public static BitVector binaryQuant(byte[] vector)
      Quantize an Int8 vector to a bit vector.
      Parameters:
      vector - the Int8 vector to quantize
      Returns:
      the result of the binary quantization of the Int8 vector

    • hammingDistance

      public static int hammingDistance(Vector<BitSet> x, Vector<BitSet> y)
      Calculate the hamming distance between two bit vectors.
      Parameters:
      x - the first bit vector
      y - the second bit vector
      Returns:
      the hamming distance between the two vectors

    • hammingDistance

      public static int hammingDistance(BitSet x, BitSet y)
      Calculate the hamming distance between two bit vectors.
      Parameters:
      x - the first bit vector
      y - the second bit vector
      Returns:
      the hamming distance between the two vectors