org.apache.spark.mllib.linalg

Class SparseMatrix

java.lang.Object

org.apache.spark.mllib.linalg.SparseMatrix

All Implemented Interfaces:

java.io.Serializable, Matrix

public class SparseMatrix
extends java.lang.Object
implements Matrix

Column-major sparse matrix. The entry values are stored in Compressed Sparse Column (CSC) format. For example, the following matrix

   1.0 0.0 4.0
   0.0 3.0 5.0
   2.0 0.0 6.0
 

is stored as values: [1.0, 2.0, 3.0, 4.0, 5.0, 6.0], rowIndices=[0, 2, 1, 0, 1, 2], colPointers=[0, 2, 3, 6].

param: numRows number of rows param: numCols number of columns param: colPtrs the index corresponding to the start of a new column (if not transposed) param: rowIndices the row index of the entry (if not transposed). They must be in strictly increasing order for each column param: values nonzero matrix entries in column major (if not transposed) param: isTransposed whether the matrix is transposed. If true, the matrix can be considered Compressed Sparse Row (CSR) format, where colPtrs behaves as rowPtrs, and rowIndices behave as colIndices, and values are stored in row major.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
SparseMatrix(int numRows, int numCols, int[] colPtrs, int[] rowIndices, double[] values) Column-major sparse matrix.
SparseMatrix(int numRows, int numCols, int[] colPtrs, int[] rowIndices, double[] values, boolean isTransposed)

Method Summary

Methods

Modifier and Type	Method and Description
double	apply(int i, int j) Gets the (i, j)-th element.
int[]	colPtrs()
SparseMatrix	copy() Get a deep copy of the matrix.
boolean	equals(java.lang.Object o)
static SparseMatrix	fromCOO(int numRows, int numCols, scala.collection.Iterable<scala.Tuple3<java.lang.Object,java.lang.Object,java.lang.Object>> entries) Generate a SparseMatrix from Coordinate List (COO) format.
boolean	isTransposed() Flag that keeps track whether the matrix is transposed or not.
int	numActives() Find the number of values stored explicitly.
int	numCols() Number of columns.
int	numNonzeros() Find the number of non-zero active values.
int	numRows() Number of rows.
int[]	rowIndices()
static SparseMatrix	spdiag(Vector vector) Generate a diagonal matrix in SparseMatrix format from the supplied values.
static SparseMatrix	speye(int n) Generate an Identity Matrix in SparseMatrix format.
static SparseMatrix	sprand(int numRows, int numCols, double density, java.util.Random rng) Generate a SparseMatrix consisting of i.i.d.
static SparseMatrix	sprandn(int numRows, int numCols, double density, java.util.Random rng) Generate a SparseMatrix consisting of i.i.d.
DenseMatrix	toDense() Generate a DenseMatrix from the given SparseMatrix.
SparseMatrix	transpose() Transpose the Matrix.
double[]	values()

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.spark.mllib.linalg.Matrix

multiply, multiply, multiply, toArray, toString, toString

Constructor Detail

SparseMatrix

public SparseMatrix(int numRows,
            int numCols,
            int[] colPtrs,
            int[] rowIndices,
            double[] values,
            boolean isTransposed)

SparseMatrix

public SparseMatrix(int numRows,
            int numCols,
            int[] colPtrs,
            int[] rowIndices,
            double[] values)

Column-major sparse matrix. The entry values are stored in Compressed Sparse Column (CSC) format. For example, the following matrix

   1.0 0.0 4.0
   0.0 3.0 5.0
   2.0 0.0 6.0
 

is stored as values: [1.0, 2.0, 3.0, 4.0, 5.0, 6.0], rowIndices=[0, 2, 1, 0, 1, 2], colPointers=[0, 2, 3, 6].

Parameters:

numRows - number of rows

numCols - number of columns

colPtrs - the index corresponding to the start of a new column

rowIndices - the row index of the entry. They must be in strictly increasing order for each column

values - non-zero matrix entries in column major

Method Detail

fromCOO

public static SparseMatrix fromCOO(int numRows,
                   int numCols,
                   scala.collection.Iterable<scala.Tuple3<java.lang.Object,java.lang.Object,java.lang.Object>> entries)

Generate a SparseMatrix from Coordinate List (COO) format. Input must be an array of (i, j, value) tuples. Entries that have duplicate values of i and j are added together. Tuples where value is equal to zero will be omitted.

Parameters:

numRows - number of rows of the matrix

numCols - number of columns of the matrix

entries - Array of (i, j, value) tuples

Returns:

The corresponding SparseMatrix

speye

public static SparseMatrix speye(int n)

Generate an Identity Matrix in SparseMatrix format.

Parameters:

n - number of rows and columns of the matrix

Returns:

SparseMatrix with size n x n and values of ones on the diagonal

sprand

public static SparseMatrix sprand(int numRows,
                  int numCols,
                  double density,
                  java.util.Random rng)

Generate a SparseMatrix consisting of i.i.d. uniform random numbers. The number of non-zero elements equal the ceiling of numRows x numCols x density

Parameters:

numRows - number of rows of the matrix

numCols - number of columns of the matrix

density - the desired density for the matrix

rng - a random number generator

Returns:

SparseMatrix with size numRows x numCols and values in U(0, 1)

sprandn

public static SparseMatrix sprandn(int numRows,
                   int numCols,
                   double density,
                   java.util.Random rng)

Generate a SparseMatrix consisting of i.i.d. gaussian random numbers.

Parameters:

numRows - number of rows of the matrix

numCols - number of columns of the matrix

density - the desired density for the matrix

rng - a random number generator

Returns:

SparseMatrix with size numRows x numCols and values in N(0, 1)

spdiag

public static SparseMatrix spdiag(Vector vector)

Generate a diagonal matrix in SparseMatrix format from the supplied values.

Parameters:

vector - a Vector that will form the values on the diagonal of the matrix

Returns:

Square SparseMatrix with size values.length x values.length and non-zero values on the diagonal

numRows

public int numRows()

Description copied from interface: Matrix

Number of rows.

Specified by:

numRows in interface Matrix

numCols

public int numCols()

Description copied from interface: Matrix

Number of columns.

Specified by:

numCols in interface Matrix

colPtrs

public int[] colPtrs()

rowIndices

public int[] rowIndices()

values

public double[] values()

isTransposed

public boolean isTransposed()

Description copied from interface: Matrix

Flag that keeps track whether the matrix is transposed or not. False by default.

Specified by:

isTransposed in interface Matrix

equals

public boolean equals(java.lang.Object o)

Overrides:

equals in class java.lang.Object

apply

public double apply(int i,
           int j)

Description copied from interface: Matrix

Gets the (i, j)-th element.

Specified by:

apply in interface Matrix

copy

public SparseMatrix copy()

Description copied from interface: Matrix

Get a deep copy of the matrix.

Specified by:

copy in interface Matrix

transpose

public SparseMatrix transpose()

Description copied from interface: Matrix

Transpose the Matrix. Returns a new `Matrix` instance sharing the same underlying data.

Specified by:

transpose in interface Matrix

toDense

public DenseMatrix toDense()

Generate a DenseMatrix from the given SparseMatrix. The new matrix will have isTransposed set to false.

Returns:

(undocumented)

numNonzeros

public int numNonzeros()

Description copied from interface: Matrix

Find the number of non-zero active values.

Specified by:

numNonzeros in interface Matrix

Returns:

(undocumented)

numActives

public int numActives()

Description copied from interface: Matrix

Find the number of values stored explicitly. These values can be zero as well.

Specified by:

numActives in interface Matrix

Returns:

(undocumented)