1 /* 2 * Licensed to the Apache Software Foundation (ASF) under one or more 3 * contributor license agreements. See the NOTICE file distributed with 4 * this work for additional information regarding copyright ownership. 5 * The ASF licenses this file to You under the Apache License, Version 2.0 6 * (the "License"); you may not use this file except in compliance with 7 * the License. You may obtain a copy of the License at 8 * 9 * https://www.apache.org/licenses/LICENSE-2.0 10 * 11 * Unless required by applicable law or agreed to in writing, software 12 * distributed under the License is distributed on an "AS IS" BASIS, 13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 * See the License for the specific language governing permissions and 15 * limitations under the License. 16 */ 17 18 /* 19 * This is not the original file distributed by the Apache Software Foundation 20 * It has been modified by the Hipparchus project 21 */ 22 23 package org.hipparchus.linear; 24 25 import org.hipparchus.exception.LocalizedCoreFormats; 26 import org.hipparchus.exception.MathIllegalArgumentException; 27 import org.hipparchus.util.FastMath; 28 29 /** 30 * Calculates the rectangular Cholesky decomposition of a matrix. 31 * <p>The rectangular Cholesky decomposition of a real symmetric positive 32 * semidefinite matrix A consists of a rectangular matrix B with the same 33 * number of rows such that: A is almost equal to BB<sup>T</sup>, depending 34 * on a user-defined tolerance. In a sense, this is the square root of A.</p> 35 * <p>The difference with respect to the regular {@link CholeskyDecomposition} 36 * is that rows/columns may be permuted (hence the rectangular shape instead 37 * of the traditional triangular shape) and there is a threshold to ignore 38 * small diagonal elements. This is used for example to generate {@link 39 * org.hipparchus.random.CorrelatedRandomVectorGenerator correlated 40 * random n-dimensions vectors} in a p-dimension subspace (p < n). 41 * In other words, it allows generating random vectors from a covariance 42 * matrix that is only positive semidefinite, and not positive definite.</p> 43 * <p>Rectangular Cholesky decomposition is <em>not</em> suited for solving 44 * linear systems, so it does not provide any {@link DecompositionSolver 45 * decomposition solver}.</p> 46 * 47 * @see <a href="http://mathworld.wolfram.com/CholeskyDecomposition.html">MathWorld</a> 48 * @see <a href="http://en.wikipedia.org/wiki/Cholesky_decomposition">Wikipedia</a> 49 */ 50 public class RectangularCholeskyDecomposition { 51 52 /** Permutated Cholesky root of the symmetric positive semidefinite matrix. */ 53 private final RealMatrix root; 54 55 /** Rank of the symmetric positive semidefinite matrix. */ 56 private int rank; 57 58 /** 59 * Decompose a symmetric positive semidefinite matrix. 60 * <p> 61 * <b>Note:</b> this constructor follows the linpack method to detect dependent 62 * columns by proceeding with the Cholesky algorithm until a nonpositive diagonal 63 * element is encountered. 64 * 65 * @see <a href="http://eprints.ma.man.ac.uk/1193/01/covered/MIMS_ep2008_56.pdf"> 66 * Analysis of the Cholesky Decomposition of a Semi-definite Matrix</a> 67 * 68 * @param matrix Symmetric positive semidefinite matrix. 69 * @exception MathIllegalArgumentException if the matrix is not 70 * positive semidefinite. 71 */ 72 public RectangularCholeskyDecomposition(RealMatrix matrix) 73 throws MathIllegalArgumentException { 74 this(matrix, 0); 75 } 76 77 /** 78 * Decompose a symmetric positive semidefinite matrix. 79 * 80 * @param matrix Symmetric positive semidefinite matrix. 81 * @param small Diagonal elements threshold under which columns are 82 * considered to be dependent on previous ones and are discarded. 83 * @exception MathIllegalArgumentException if the matrix is not 84 * positive semidefinite. 85 */ 86 public RectangularCholeskyDecomposition(RealMatrix matrix, double small) 87 throws MathIllegalArgumentException { 88 89 final int order = matrix.getRowDimension(); 90 final double[][] c = matrix.getData(); 91 final double[][] b = new double[order][order]; 92 93 int[] index = new int[order]; 94 for (int i = 0; i < order; ++i) { 95 index[i] = i; 96 } 97 98 int r = 0; 99 for (boolean loop = true; loop;) { 100 101 // find maximal diagonal element 102 int swapR = r; 103 for (int i = r + 1; i < order; ++i) { 104 int ii = index[i]; 105 int isr = index[swapR]; 106 if (c[ii][ii] > c[isr][isr]) { 107 swapR = i; 108 } 109 } 110 111 112 // swap elements 113 if (swapR != r) { 114 final int tmpIndex = index[r]; 115 index[r] = index[swapR]; 116 index[swapR] = tmpIndex; 117 final double[] tmpRow = b[r]; 118 b[r] = b[swapR]; 119 b[swapR] = tmpRow; 120 } 121 122 // check diagonal element 123 int ir = index[r]; 124 if (c[ir][ir] <= small) { 125 126 if (r == 0) { 127 throw new MathIllegalArgumentException(LocalizedCoreFormats.NOT_POSITIVE_DEFINITE_MATRIX); 128 } 129 130 // check remaining diagonal elements 131 for (int i = r; i < order; ++i) { 132 if (c[index[i]][index[i]] < -small) { 133 // there is at least one sufficiently negative diagonal element, 134 // the symmetric positive semidefinite matrix is wrong 135 throw new MathIllegalArgumentException(LocalizedCoreFormats.NOT_POSITIVE_DEFINITE_MATRIX); 136 } 137 } 138 139 // all remaining diagonal elements are close to zero, we consider we have 140 // found the rank of the symmetric positive semidefinite matrix 141 loop = false; 142 143 } else { 144 145 // transform the matrix 146 final double sqrt = FastMath.sqrt(c[ir][ir]); 147 b[r][r] = sqrt; 148 final double inverse = 1 / sqrt; 149 final double inverse2 = 1 / c[ir][ir]; 150 for (int i = r + 1; i < order; ++i) { 151 final int ii = index[i]; 152 final double e = inverse * c[ii][ir]; 153 b[i][r] = e; 154 c[ii][ii] -= c[ii][ir] * c[ii][ir] * inverse2; 155 for (int j = r + 1; j < i; ++j) { 156 final int ij = index[j]; 157 final double f = c[ii][ij] - e * b[j][r]; 158 c[ii][ij] = f; 159 c[ij][ii] = f; 160 } 161 } 162 163 // prepare next iteration 164 loop = ++r < order; 165 } 166 } 167 168 // build the root matrix 169 rank = r; 170 root = MatrixUtils.createRealMatrix(order, r); 171 for (int i = 0; i < order; ++i) { 172 for (int j = 0; j < r; ++j) { 173 root.setEntry(index[i], j, b[i][j]); 174 } 175 } 176 177 } 178 179 /** Get the root of the covariance matrix. 180 * The root is the rectangular matrix <code>B</code> such that 181 * the covariance matrix is equal to <code>B.B<sup>T</sup></code> 182 * @return root of the square matrix 183 * @see #getRank() 184 */ 185 public RealMatrix getRootMatrix() { 186 return root; 187 } 188 189 /** Get the rank of the symmetric positive semidefinite matrix. 190 * The r is the number of independent rows in the symmetric positive semidefinite 191 * matrix, it is also the number of columns of the rectangular 192 * matrix of the decomposition. 193 * @return r of the square matrix. 194 * @see #getRootMatrix() 195 */ 196 public int getRank() { 197 return rank; 198 } 199 200 }