Package org.ejml.ops

Source Code of org.ejml.ops.TestRandomMatrices

/*
* Copyright (c) 2009-2012, Peter Abeles. All Rights Reserved.
*
* This file is part of Efficient Java Matrix Library (EJML).
*
* EJML is free software: you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation, either version 3
* of the License, or (at your option) any later version.
*
* EJML is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with EJML.  If not, see <http://www.gnu.org/licenses/>.
*/

package org.ejml.ops;

import org.ejml.alg.dense.mult.VectorVectorMult;
import org.ejml.data.Complex64F;
import org.ejml.data.DenseMatrix64F;
import org.ejml.data.UtilTestMatrix;
import org.ejml.factory.DecompositionFactory;
import org.ejml.factory.EigenDecomposition;
import org.ejml.factory.SingularValueDecomposition;
import org.junit.Test;

import java.util.Arrays;
import java.util.Random;

import static org.junit.Assert.*;


/**
* @author Peter Abeles
*/
public class TestRandomMatrices {

    Random rand = new Random(48757);

    /**
     * Checks to see if all the vectors are orthogonal and of unit length.
     */
    @Test
    public void createSpan() {
        // test with combinations of vectors and numbers
        for( int dimension = 3; dimension <= 5; dimension++ ) {
            for( int numVectors = 1; numVectors <= dimension; numVectors++ ) {
                DenseMatrix64F span[] = RandomMatrices.createSpan(dimension,numVectors,rand);

                assertEquals(numVectors,span.length);

                for( int i = 0; i < span.length; i++ ) {
                    DenseMatrix64F a = span[i];

                    assertEquals(1,NormOps.fastNormF(a),1e-8);

                    for( int j = i+1; j < span.length; j++ ) {
                        double dot = VectorVectorMult.innerProd(a,span[j]);
                        assertEquals(0,dot,1e-8);
                    }
                }
            }
        }
    }

    @Test
    public void createInSpan() {
        DenseMatrix64F span[] = RandomMatrices.createSpan(5,5,rand);

        DenseMatrix64F A = RandomMatrices.createInSpan(span,-1,1,rand);

        // reconstructed matrix
        DenseMatrix64F R = new DenseMatrix64F(A.numRows,A.numCols);

        DenseMatrix64F tmp = new DenseMatrix64F(A.numRows,A.numCols);

        // project the matrix into the span and recreate the original matrix
        for( int i = 0; i < 5; i++ ) {
            double val =  VectorVectorMult.innerProd(span[i],A);
            assertTrue( Math.abs(val) > 1e-10 );

            CommonOps.scale(val,span[i],tmp);
            CommonOps.add(R,tmp,R);
        }

        double error = SpecializedOps.diffNormF(A,R);

        assertEquals( error , 0 , 1e-8 );
    }

    @Test
    public void createOrthogonal() {
        for( int numRows = 3; numRows <= 5; numRows++ ) {
            for( int numCols = 1; numCols <= numRows; numCols++ ) {
                DenseMatrix64F Q = RandomMatrices.createOrthogonal(numRows,numCols,rand);

                assertEquals(Q.numRows,numRows);
                assertEquals(Q.numCols,numCols);

                assertTrue(CommonOps.elementSum(Q) != 0);
                assertTrue(MatrixFeatures.isOrthogonal(Q,1e-8));
            }
        }
    }

    @Test
    public void createDiagonal_square() {
        DenseMatrix64F A = RandomMatrices.createDiagonal(5,1,10,rand);

        assertTrue(CommonOps.elementSum(A) > 5 );
        for( int i = 0; i < A.numRows; i++ ) {
            for( int j = 0; j < A.numCols; j++ ) {
                double v = A.get(i,j);

                if( i == j ) {
                    assertTrue(v >= 1 || v <= 10);
                } else {
                    assertTrue(v == 0);
                }
            }
        }
    }

    @Test
    public void createDiagonal_general() {
        testDiagonal(5,3);
        testDiagonal(3,5);
        testDiagonal(3,3);
    }

    public void testDiagonal( int numRows , int numCols ) {
        DenseMatrix64F A = RandomMatrices.createDiagonal(numRows,numCols,1,10,rand);

        assertEquals(A.getNumRows(),numRows);
        assertEquals(A.getNumCols(),numCols);

        assertTrue(CommonOps.elementSum(A) > 5 );
        for( int i = 0; i < A.numRows; i++ ) {
            for( int j = 0; j < A.numCols; j++ ) {
                double v = A.get(i,j);

                if( i == j ) {
                    assertTrue(v >= 1 || v <= 10);
                } else {
                    assertTrue(v == 0);
                }
            }
        }
    }

    @Test
    public void createSingularValues() {
        // check case when sv is more than or equal to the matrix dimension
        double sv[] = new double[]{8.2,6.2,4.1,2};

        for( int numRows = 1; numRows <= 4; numRows++ ) {
            for( int numCols = 1; numCols <= 4; numCols++ ) {
                DenseMatrix64F A = RandomMatrices.createSingularValues(numRows,numCols, rand, sv);

                SingularValueDecomposition<DenseMatrix64F> svd =
                        DecompositionFactory.svd(A.numRows,A.numCols,true,true,false);
                assertTrue(svd.decompose(A));

                int o = Math.min(numRows,numCols);

                UtilTestMatrix.checkSameElements(1e-8,o,sv,svd.getSingularValues());
            }
        }

        // see if it fills in zeros when it is smaller than the dimension
        DenseMatrix64F A = RandomMatrices.createSingularValues(5,5, rand, sv);

        SingularValueDecomposition<DenseMatrix64F> svd =
                DecompositionFactory.svd(A.numRows,A.numCols,true,true,false);
        assertTrue(svd.decompose(A));

        UtilTestMatrix.checkSameElements(1e-8,sv.length,sv,svd.getSingularValues());
        assertEquals(0,svd.getSingularValues()[4],1e-8);
    }

    @Test
    public void createEigenvaluesSymm() {
        DenseMatrix64F A = RandomMatrices.createEigenvaluesSymm(5,rand,1,2,3,4,5);

        // this should be symmetric
        assertTrue(MatrixFeatures.isSymmetric(A,1e-10));

        // decompose the matrix and extract its eigenvalues
        EigenDecomposition<DenseMatrix64F> eig = DecompositionFactory.eig(A.numRows,true);
        assertTrue(eig.decompose(A));

        double ev[] = new double[5];
        for( int i = 0; i < 5; i++ ) {
            Complex64F e = eig.getEigenvalue(i);
            assertTrue(e.isReal());

            ev[i] = e.real;
        }

        // need to sort the eigenvalues so that I know where they are in the array
        Arrays.sort(ev);

        // see if they are what I expected them to be
        for( int i = 0; i < ev.length; i++ ) {
            assertEquals(i+1.0,ev[i],1e-8);
        }
    }

    @Test
    public void addRandom() {
        DenseMatrix64F A = new DenseMatrix64F(3,4);

        CommonOps.fill(A, -2.0);

        RandomMatrices.addRandom(A,1,2,rand);

        for( int i = 0; i < A.getNumElements(); i++ ) {
            assertTrue(A.get(i) >= -1 && A.get(i) <= 0 );
        }
    }

    @Test
    public void createRandom() {
        DenseMatrix64F A = RandomMatrices.createRandom(5,4,rand);

        checkRandom1(A);
    }

    @Test
    public void createRandom_min_max() {
        DenseMatrix64F A = RandomMatrices.createRandom(30,20,-1,1,rand);

        checkRandomRange(A);
    }

    @Test
    public void setRandom() {
        DenseMatrix64F A = new DenseMatrix64F(5,4);

        RandomMatrices.setRandom(A,rand);

        checkRandom1(A);
    }

    private void checkRandom1(DenseMatrix64F a) {
        assertEquals(5, a.numRows);
        assertEquals(4, a.numCols);

        double total = 0;
        for( int i = 0; i < a.numRows; i++ ) {
            for( int j = 0; j < a.numCols; j++ ) {
                double val = a.get(i,j);

                assertTrue( val >= 0);
                assertTrue( val <= 1);
                total += val;
            }
        }

        assertTrue(total>0);
    }

    @Test
    public void setRandom_min_max() {
        DenseMatrix64F A = new DenseMatrix64F(30,20);
        RandomMatrices.setRandom(A,-1,1,rand);

        checkRandomRange(A);
    }

    private void checkRandomRange(DenseMatrix64F a) {
        assertEquals(30, a.numRows);
        assertEquals(20, a.numCols);

        int numNeg = 0;
        int numPos = 0;
        for( int i = 0; i < a.numRows; i++ ) {
            for( int j = 0; j < a.numCols; j++ ) {
                double val = a.get(i,j);

                if( val < 0 )
                    numNeg++;
                else
                    numPos++;

                if( Math.abs(val) > 1 )
                    fail("Out of range");
            }
        }

        assertTrue(numNeg>0);
        assertTrue(numPos>0);
    }

    @Test
    public void createSymmPosDef() {
        for( int i = 0; i < 10; i++ ) {
            DenseMatrix64F A = RandomMatrices.createSymmPosDef(6+i,rand);

            assertTrue(MatrixFeatures.isPositiveDefinite(A));
        }
    }

    @Test
    public void createSymmetric() {
        DenseMatrix64F A = RandomMatrices.createSymmetric(10,-1,1,rand);

        assertTrue(MatrixFeatures.isSymmetric(A,1e-8));

        // see if it has the expected range of elements
        double min = CommonOps.elementMin(A);
        double max = CommonOps.elementMax(A);

        assertTrue(min < 0 && min >= -1);
        assertTrue(max > 0 && max <= 1);
    }

    @Test
    public void createUpperTriangle() {
        for( int hess = 0; hess < 3; hess++ ) {
            DenseMatrix64F A = RandomMatrices.createUpperTriangle(10,hess,-1,1,rand);

            assertTrue(MatrixFeatures.isUpperTriangle(A,hess,1e-8));

            // quick sanity check to make sure it could be proper
            assertTrue(A.get(hess,0) != 0 );

            // see if it has the expected range of elements
            double min = CommonOps.elementMin(A);
            double max = CommonOps.elementMax(A);

            assertTrue(min < 0 && min >= -1);
            assertTrue(max > 0 && max <= 1);
        }
    }
}
TOP

Related Classes of org.ejml.ops.TestRandomMatrices

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.