LMcovar.h
Go to the documentation of this file.
00001 // This file is part of Eigen, a lightweight C++ template library
00002 // for linear algebra.
00003 //
00004 // This code initially comes from MINPACK whose original authors are:
00005 // Copyright Jorge More - Argonne National Laboratory
00006 // Copyright Burt Garbow - Argonne National Laboratory
00007 // Copyright Ken Hillstrom - Argonne National Laboratory
00008 //
00009 // This Source Code Form is subject to the terms of the Minpack license
00010 // (a BSD-like license) described in the campaigned CopyrightMINPACK.txt file.
00011 
00012 #ifndef EIGEN_LMCOVAR_H
00013 #define EIGEN_LMCOVAR_H
00014 
00015 namespace Eigen { 
00016 
00017 namespace internal {
00018 
00019 template <typename Scalar>
00020 void covar(
00021         Matrix< Scalar, Dynamic, Dynamic > &r,
00022         const VectorXi& ipvt,
00023         Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) )
00024 {
00025     using std::abs;
00026     typedef DenseIndex Index;
00027     /* Local variables */
00028     Index i, j, k, l, ii, jj;
00029     bool sing;
00030     Scalar temp;
00031 
00032     /* Function Body */
00033     const Index n = r.cols();
00034     const Scalar tolr = tol * abs(r(0,0));
00035     Matrix< Scalar, Dynamic, 1 > wa(n);
00036     eigen_assert(ipvt.size()==n);
00037 
00038     /* form the inverse of r in the full upper triangle of r. */
00039     l = -1;
00040     for (k = 0; k < n; ++k)
00041         if (abs(r(k,k)) > tolr) {
00042             r(k,k) = 1. / r(k,k);
00043             for (j = 0; j <= k-1; ++j) {
00044                 temp = r(k,k) * r(j,k);
00045                 r(j,k) = 0.;
00046                 r.col(k).head(j+1) -= r.col(j).head(j+1) * temp;
00047             }
00048             l = k;
00049         }
00050 
00051     /* form the full upper triangle of the inverse of (r transpose)*r */
00052     /* in the full upper triangle of r. */
00053     for (k = 0; k <= l; ++k) {
00054         for (j = 0; j <= k-1; ++j)
00055             r.col(j).head(j+1) += r.col(k).head(j+1) * r(j,k);
00056         r.col(k).head(k+1) *= r(k,k);
00057     }
00058 
00059     /* form the full lower triangle of the covariance matrix */
00060     /* in the strict lower triangle of r and in wa. */
00061     for (j = 0; j < n; ++j) {
00062         jj = ipvt[j];
00063         sing = j > l;
00064         for (i = 0; i <= j; ++i) {
00065             if (sing)
00066                 r(i,j) = 0.;
00067             ii = ipvt[i];
00068             if (ii > jj)
00069                 r(ii,jj) = r(i,j);
00070             if (ii < jj)
00071                 r(jj,ii) = r(i,j);
00072         }
00073         wa[jj] = r(j,j);
00074     }
00075 
00076     /* symmetrize the covariance matrix in r. */
00077     r.topLeftCorner(n,n).template triangularView<StrictlyUpper>() = r.topLeftCorner(n,n).transpose();
00078     r.diagonal() = wa;
00079 }
00080 
00081 } // end namespace internal
00082 
00083 } // end namespace Eigen
00084 
00085 #endif // EIGEN_LMCOVAR_H


shape_reconstruction
Author(s): Roberto Martín-Martín
autogenerated on Sat Jun 8 2019 18:32:50