/*
* 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.data.DenseMatrix64F;
import org.ejml.simple.SimpleMatrix;
import org.junit.Test;
import java.util.Random;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertFalse;
/**
* @author Peter Abeles
*/
public class TestNormOps {
Random rand = new Random(234);
DenseMatrix64F zeroMatrix = new DenseMatrix64F(3,4);
DenseMatrix64F unzeroMatrix = new DenseMatrix64F(3,2, true, 0.2, 1, -2, 3, 6, 5);
DenseMatrix64F unzeroVector = new DenseMatrix64F(5,1, true, 0.3, 1, -2, 3, 4);
DenseMatrix64F squareMatrix = new DenseMatrix64F(2,2, true, 0.2, 1, -2, 3);
/**
* Tests against the condition number from octave.
*/
@Test
public void conditionP() {
double val = NormOps.conditionP(squareMatrix,1);
assertEquals(7.6923,val,1e-3);
// check the non-square case
val = NormOps.conditionP(unzeroMatrix,1);
assertEquals(3.4325,val,1e-3);
// see if the other pseudo-inverse works
DenseMatrix64F trans = unzeroMatrix.copy();
CommonOps.transpose(trans);
val = NormOps.conditionP(trans,1);
assertEquals(3.4887,val,1e-3);
}
/**
* Tests against the condition number from octave.
*/
@Test
public void conditionP2() {
double val = NormOps.conditionP2(unzeroMatrix);
assertEquals(2.1655,val,1e-3);
checkUncountable(NormOps.conditionP2(zeroMatrix));
}
/**
* Tested using the following operation in octave:
*
* sum(abs(a(:)).^3.5)^(1/3.5)
*/
@Test
public void elementP() {
double val = NormOps.elementP(unzeroMatrix,3.5);
assertEquals(6.9108,val,1e-3);
checkUncountable(NormOps.elementP(zeroMatrix,3.5));
}
@Test
public void fastElementP() {
double val = NormOps.fastElementP(unzeroMatrix,3.5);
assertEquals(6.9108,val,1e-3);
checkUncountable(NormOps.fastElementP(zeroMatrix,3.5));
}
@Test
public void normalizeF() {
DenseMatrix64F a = unzeroVector.copy();
NormOps.normalizeF(a);
assertEquals(1,NormOps.normF(a),1e-6);
}
@Test
public void fastNormF() {
double val = NormOps.fastNormF(unzeroMatrix);
assertEquals(8.6626,val,1e-3);
checkUncountable(NormOps.fastNormF(zeroMatrix));
}
@Test
public void normF() {
double val = NormOps.normF(unzeroMatrix);
assertEquals(8.6626,val,1e-3);
checkUncountable(NormOps.normF(zeroMatrix));
}
@Test
public void fastNormP2() {
// check induced matrix norm
double found = NormOps.fastNormP2(unzeroMatrix);
double expected = NormOps.inducedP2(unzeroMatrix);
assertEquals(expected,found,1e-3);
// check vector norm
found = NormOps.fastNormP2(unzeroVector);
expected = NormOps.normF(unzeroVector);
assertEquals(expected,found,1e-3);
}
@Test
public void normP() {
// check induced matrix norm
double found = NormOps.normP(unzeroMatrix,2);
double expected = NormOps.inducedP2(unzeroMatrix);
assertEquals(expected,found,1e-3);
// check vector norm
found = NormOps.normP(unzeroVector,2);
expected = NormOps.normF(unzeroVector);
assertEquals(expected,found,1e-3);
}
@Test
public void fastNormP() {
// check induced matrix norm
double found = NormOps.fastNormP(unzeroMatrix,2);
double expected = NormOps.inducedP2(unzeroMatrix);
assertEquals(expected,found,1e-3);
// check vector norm
found = NormOps.fastNormP(unzeroVector,2);
expected = NormOps.normF(unzeroVector);
assertEquals(expected,found,1e-3);
}
@Test
public void normP1() {
// check induced matrix norm
double found = NormOps.normP1(unzeroMatrix);
double expected = NormOps.inducedP1(unzeroMatrix);
assertEquals(expected,found,1e-3);
// check vector norm
found = NormOps.normP1(unzeroVector);
expected = CommonOps.elementSumAbs(unzeroVector);
assertEquals(expected,found,1e-3);
}
@Test
public void normP2() {
// check induced matrix norm
double found = NormOps.normP2(unzeroMatrix);
double expected = NormOps.inducedP2(unzeroMatrix);
assertEquals(expected,found,1e-3);
// check vector norm
found = NormOps.normP2(unzeroVector);
expected = NormOps.normF(unzeroVector);
assertEquals(expected,found,1e-3);
}
@Test
public void normPInf() {
// check induced matrix norm
double found = NormOps.normPInf(unzeroMatrix);
double expected = NormOps.inducedPInf(unzeroMatrix);
assertEquals(expected,found,1e-3);
// check vector norm
found = NormOps.normPInf(unzeroVector);
expected = CommonOps.elementMaxAbs(unzeroVector);
assertEquals(expected,found,1e-3);
}
@Test
public void inducedP1() {
double val = NormOps.inducedP1(unzeroMatrix);
assertEquals(9,val,1e-3);
checkUncountable(NormOps.inducedP1(zeroMatrix));
}
@Test
public void inducedP2() {
double val = NormOps.inducedP2(unzeroMatrix);
assertEquals(7.8645,val,1e-3);
checkUncountable(NormOps.inducedP2(zeroMatrix));
// make sure the largest singular value is being returned not just the first
for( int i = 0; i < 20; i++ ) {
SimpleMatrix A = SimpleMatrix.random(5,5,-10,10,rand);
double largest = A.svd().getW().get(0);
assertEquals(largest,NormOps.inducedP2(A.getMatrix()),1e-8);
}
}
@Test
public void inducedPInf() {
double val = NormOps.inducedPInf(unzeroMatrix);
assertEquals(11,val,1e-3);
checkUncountable(NormOps.inducedPInf(zeroMatrix));
}
private static void checkUncountable( double val ) {
assertFalse(Double.isInfinite(val));
assertFalse(Double.isNaN(val));
}
}