Class ApproximateSparseVector

A one-dimensional vector of double values, optimised for the case where many of the elements share a common value (which need not be zero) within some tolerance.

Inheritance
Object
Vector
SparseVector
ApproximateSparseVector
Inherited Members
Namespace: Microsoft.ML.Probabilistic.Math
Assembly: Microsoft.ML.Probabilistic.dll
Syntax
[Serializable]
[DataContract]
public class ApproximateSparseVector : SparseVector, IReadOnlyList<double>, IReadOnlyCollection<double>, SettableToPower<Vector>, SettableToProduct<Vector>, SettableToProduct<Vector, Vector>, SettableToWeightedSum<Vector>, ISparseList<double>, ISparseEnumerable<double>, IList<double>, ICollection<double>, IEnumerable<double>, IEnumerable, ICloneable, CanSetAllElementsTo<double>, SettableTo<Vector>

Properties

CountTolerance

The maximum allowed count of vector elements not set to the common value.

Declaration
public int CountTolerance { get; }
Property Value
Type Description
Int32
Remarks

This is ignored if <= 0 which is the default value.

Item[Int32]

Gets or sets an element at a given index.

Declaration
public override double this[int index] { get; set; }
Parameters
Type Name Description
Int32 index

The index of an element.
Property Value
Type Description
Double

The element at a given index.
Overrides

Tolerance

The tolerance at which vector element values are considered equal to the common value.

Declaration
public double Tolerance { get; }
Property Value
Type Description
Double
Remarks

By default this tolerance is set to 0.000001.

Methods

Append(Vector)

Returns a new vector which appends a second vector to this vector

Declaration
public override Vector Append(Vector second)
Parameters
Type Name Description
Vector second

Second vector
Returns
Type Description
Vector
Overrides
Remarks

If the second vector is dense, or if the common values for the two vectors are different, then the result becomes dense

Append(Double)

Appends an item to a vector - returns a new sparse vector

Declaration
public override Vector Append(double item)
Parameters
Type Name Description
Double item
Returns
Type Description
Vector
Overrides

Clone()

Clones this vector - return as a vector

Declaration
public override Vector Clone()
Returns
Type Description
Vector
Overrides

Constant(Int32, Double)

Creates an approximate sparse vector of given length with elements all equal to a specified value

Declaration
public static ApproximateSparseVector Constant(int count, double value)
Parameters
Type Name Description
Int32 count

Number of elements in vector
Double value

value for each element
Returns
Type Description
ApproximateSparseVector

Constant(Int32, Double, Sparsity)

Creates an approximate sparse vector of given length with elements all equal to a specified value

Declaration
public static ApproximateSparseVector Constant(int count, double value, Sparsity s)
Parameters
Type Name Description
Int32 count

Number of elements in vector
Double value

value for each element
Sparsity s

Sparsity
Returns
Type Description
ApproximateSparseVector

Contains(Double)

Returns true if the Vector contains the specified value up to tolerance

Declaration
public override bool Contains(double value)
Parameters
Type Name Description
Double value

The value to test for
Returns
Type Description
Boolean
Overrides

Copy(Vector)

Creates an approximate sparse vector as a copy of another vector (which may not be sparse)

Declaration
public static ApproximateSparseVector Copy(Vector that)
Parameters
Type Name Description
Vector that

The source vector - can be dense or sparse
Returns
Type Description
ApproximateSparseVector

FromArray(Double[])

Constructs a sparse vector from a dense array.

Declaration
public static ApproximateSparseVector FromArray(params double[] data)
Parameters
Type Name Description
Double[] data

1D array of elements.
Returns
Type Description
ApproximateSparseVector
Remarks

The array data is copied into new storage. The size of the vector is taken from the array.

FromArray(Int32, Double[], Int32)

Constructs a vector from part of an array.

Declaration
public static ApproximateSparseVector FromArray(int count, double[] data, int start)
Parameters
Type Name Description
Int32 count

The number of elements in the vector.
Double[] data

Storage for the vector elements.
Int32 start

The starting index in the array for the vector elements.
Returns
Type Description
ApproximateSparseVector
Remarks

Throws an exception if Data is null, start < 0, or count < 0.

FromSparseValues(Int32, Double, Sparsity, List<ValueAtIndex<Double>>)

Constructs a sparse vector from a sorted list of sparse elements.

Declaration
[Construction(new string[]{"Count", "CommonValue", "Sparsity", "SparseValues"})]
public static ApproximateSparseVector FromSparseValues(int count, double commonValue, Sparsity s, List<ValueAtIndex<double>> sortedSparseValues)
Parameters
Type Name Description
Int32 count

Count for result
Double commonValue

Common value
Sparsity s

Sparsity
List<ValueAtIndex<Double>> sortedSparseValues

Sorted list of sparse elements
Returns
Type Description
ApproximateSparseVector

GetHashCode()

Gets a hash code for the instance.

Declaration
public override int GetHashCode()
Returns
Type Description
Int32

The code.
Overrides

IndexOf(Double)

Returns the index of the first occurence of the given value in the array. Returns -1 if the value is not in the array

Declaration
public override int IndexOf(double item)
Parameters
Type Name Description
Double item

The item to check for
Returns
Type Description
Int32

Its index in the array
Overrides

MaxDiff(Vector)

Returns the maximum absolute difference between this vector and another vector.

Declaration
public override double MaxDiff(Vector that)
Parameters
Type Name Description
Vector that

The second vector.
Returns
Type Description
Double

max(abs(this[i] - that[i])). Matching infinities or NaNs do not count.
If this and that are not the same size, returns infinity.
Overrides
Remarks

This routine is typically used instead of Equals, since Equals is susceptible to roundoff errors.

MaxDiff(Vector, Double)

Returns the maximum relative difference between this vector and another.

Declaration
public override double MaxDiff(Vector that, double rel)
Parameters
Type Name Description
Vector that

The second vector.
Double rel

An offset to avoid division by zero.
Returns
Type Description
Double

max(abs(this[i] - that[i])/(abs(this[i]) + rel)). Matching infinities or NaNs do not count.
If this and that are not the same size, returns infinity.
Overrides
Remarks

This routine is typically used instead of Equals, since Equals is susceptible to roundoff errors.

Scale(Double)

Multiplies this vector by a scalar.

Declaration
public override void Scale(double scale)
Parameters
Type Name Description
Double scale

The scalar.
Overrides
Remarks

this receives the product. This method is a synonym for SetToProduct(this, scale)

SetSubvector(Int32, Vector)

Set a subvector of this sparse vector to another vector. The common value is not changed

Declaration
public override void SetSubvector(int startIndex, Vector that)
Parameters
Type Name Description
Int32 startIndex

The index of the first element of this to copy to.
Vector that

A vector whose length is at most this.Count - startIndex.
Overrides

SetTo(ApproximateSparseVector)

Copies values from a sparse vector to this sparse vector.

Declaration
public void SetTo(ApproximateSparseVector that)
Parameters
Type Name Description
ApproximateSparseVector that

SetTo(Vector)

Copies values from a Vector to this vector.

Declaration
public override void SetTo(Vector that)
Parameters
Type Name Description
Vector that
Overrides
Remarks

The source vector can be dense, in which case the minimum value is used as the common value.

SetTo(IEnumerable<Double>)

Set this vector to a collection

Declaration
public override void SetTo(IEnumerable<double> that)
Parameters
Type Name Description
IEnumerable<Double> that
Overrides

SetTo(IList<Double>, Double)

Copies values from a list of doubles which must have the same size as this vector, using the specified common value.

Declaration
public override void SetTo(IList<double> dlist, double commonValue)
Parameters
Type Name Description
IList<Double> dlist

List of doubles
Double commonValue

Common value
Overrides

SetTo(Double[])

Copies values from an array. The minimum value is used as the common value

Declaration
public override void SetTo(double[] values)
Parameters
Type Name Description
Double[] values

An array whose length matches this.Count.
Overrides

SetToDiagonal(Matrix)

Sets this vector to the diagonal of a matrix.

Declaration
public override void SetToDiagonal(Matrix m)
Parameters
Type Name Description
Matrix m

A matrix with Rows==Cols==this.Count.
Overrides

SetToFunction(SparseVector, SparseVector, Func<Double, Double, Double>)

Sets the elements of this sparse vector to a function of the elements of two other sparse vectors

Declaration
public override SparseVector SetToFunction(SparseVector a, SparseVector b, Func<double, double, double> fun)
Parameters
Type Name Description
SparseVector a

The first vector
SparseVector b

The second vector
Func<Double, Double, Double> fun

The function which maps two doubles to a double
Returns
Type Description
SparseVector
Overrides
Remarks

Assumes the vectors are compatible

SetToFunction(SparseVector, Converter<Double, Double>)

Sets the elements of this sparse vector to a function of the elements of another sparse vector

Declaration
public override SparseVector SetToFunction(SparseVector that, Converter<double, double> fun)
Parameters
Type Name Description
SparseVector that

The other vector
Converter<Double, Double> fun

The function which maps doubles to doubles
Returns
Type Description
SparseVector
Overrides
Remarks

Assumes the vectors are compatible

SetToFunctionInPlace(SparseVector, Func<Double, Double, Double>)

Sets the elements of this sparse vector to a function of the elements of this sparse vector and another sparse vectors x = fun(x,b)

Declaration
public override SparseVector SetToFunctionInPlace(SparseVector b, Func<double, double, double> fun)
Parameters
Type Name Description
SparseVector b

The second vector
Func<Double, Double, Double> fun

The function which maps two doubles to a double
Returns
Type Description
SparseVector
Overrides
Remarks

Assumes the vectors are compatible

SetToSubarray(Double[], Int32)

Copies values from an array. The minimum is used as the common value

Declaration
public override void SetToSubarray(double[] values, int startIndex)
Parameters
Type Name Description
Double[] values

An array whose length is at least this.Count + startIndex.
Int32 startIndex

The index of the first value in values to copy.
Overrides

SetToSubvector(SparseVector, Int32)

Copies values from a sparse vector. The common value is set to the common value from the source vector.

Declaration
public override void SetToSubvector(SparseVector that, int startIndex)
Parameters
Type Name Description
SparseVector that

A vector whose length is at least this.Count + startIndex.
Int32 startIndex

The index of the first value in that to copy.
Overrides

Subvector(Int32, Int32)

Create a subvector of this sparse vector.

Declaration
public override Vector Subvector(int startIndex, int count)
Parameters
Type Name Description
Int32 startIndex
Int32 count
Returns
Type Description
Vector
Overrides

ToString(String, String)

String representation of vector with a specified format and delimiter

Declaration
public override string ToString(string format, string delimiter)
Parameters
Type Name Description
String format
String delimiter
Returns
Type Description
String
Overrides

Zero(Int32)

Creates an approximate sparse vector of given length with elements all 0.0

Declaration
public static ApproximateSparseVector Zero(int count)
Parameters
Type Name Description
Int32 count

Number of elements in vector
Returns
Type Description
ApproximateSparseVector

Zero(Int32, Sparsity)

Creates an approximate sparse vector of given length with elements all 0.0

Declaration
public static ApproximateSparseVector Zero(int count, Sparsity s)
Parameters
Type Name Description
Int32 count

Number of elements in vector
Sparsity s

Sparsity
Returns
Type Description
ApproximateSparseVector

Explicit Interface Implementations

ICloneable.Clone()

Clones this vector - return as an object

Declaration
object ICloneable.Clone()
Returns
Type Description
Object

Implements

Extension Methods