概要
本サンプルはFortran言語によりLAPACKルーチンZGELSSを利用するサンプルプログラムです。
以下の線形最小二乗問題を解きます。最小ノルム解を

及び

入力データ
(本ルーチンの詳細はZGELSS のマニュアルページを参照)| このデータをダウンロード |
ZGELSS Example Program Data 5 4 :Values of M and N ( 0.47,-0.34) (-0.40, 0.54) ( 0.60, 0.01) ( 0.80,-1.02) (-0.32,-0.23) (-0.05, 0.20) (-0.26,-0.44) (-0.43, 0.17) ( 0.35,-0.60) (-0.52,-0.34) ( 0.87,-0.11) (-0.34,-0.09) ( 0.89, 0.71) (-0.45,-0.45) (-0.02,-0.57) ( 1.14,-0.78) (-0.19, 0.06) ( 0.11,-0.85) ( 1.44, 0.80) ( 0.07, 1.14) :End of matrix A (-1.08,-2.59) (-2.61,-1.49) ( 3.13,-3.61) ( 7.33,-8.01) ( 9.12, 7.63) :End of vector b
出力結果
(本ルーチンの詳細はZGELSS のマニュアルページを参照)| この出力例をダウンロード |
ZGELSS Example Program Results
Least squares solution
( 1.1673,-3.3222) ( 1.3480, 5.5028) ( 4.1762, 2.3434) ( 0.6465, 0.0105)
Tolerance used to estimate the rank of A
1.00E-02
Estimated rank of A
3
Singular values of A
2.9979 1.9983 1.0044 0.0064
ソースコード
(本ルーチンの詳細はZGELSS のマニュアルページを参照)※本サンプルソースコードのご利用手順は「サンプルのコンパイル及び実行方法」をご参照下さい。
| このソースコードをダウンロード |
Program zgelss_example
! ZGELSS Example Program Text
! Copyright 2017, Numerical Algorithms Group Ltd. http://www.nag.com
! .. Use Statements ..
Use blas_interfaces, Only: dznrm2
Use lapack_interfaces, Only: zgelss
Use lapack_precision, Only: dp
! .. Implicit None Statement ..
Implicit None
! .. Parameters ..
Integer, Parameter :: nb = 64, nin = 5, nout = 6
! .. Local Scalars ..
Real (Kind=dp) :: rcond, rnorm
Integer :: i, info, lda, lwork, m, n, rank
! .. Local Arrays ..
Complex (Kind=dp), Allocatable :: a(:, :), b(:), work(:)
Real (Kind=dp), Allocatable :: rwork(:), s(:)
! .. Executable Statements ..
Write (nout, *) 'ZGELSS Example Program Results'
Write (nout, *)
! Skip heading in data file
Read (nin, *)
Read (nin, *) m, n
lda = m
lwork = 2*n + nb*(m+n)
Allocate (a(lda,n), b(m), work(lwork), rwork(5*n), s(n))
! Read A and B from data file
Read (nin, *)(a(i,1:n), i=1, m)
Read (nin, *) b(1:m)
! Choose RCOND to reflect the relative accuracy of the input data
rcond = 0.01E0_dp
! Solve the least squares problem min( norm2(b - Ax) ) for the x
! of minimum norm.
Call zgelss(m, n, 1, a, lda, b, m, s, rcond, rank, work, lwork, rwork, &
info)
If (info==0) Then
! Print solution
Write (nout, *) 'Least squares solution'
Write (nout, 100) b(1:n)
! Print the effective rank of A
Write (nout, *)
Write (nout, *) 'Tolerance used to estimate the rank of A'
Write (nout, 110) rcond
Write (nout, *) 'Estimated rank of A'
Write (nout, 120) rank
! Print singular values of A
Write (nout, *)
Write (nout, *) 'Singular values of A'
Write (nout, 130) s(1:n)
! Compute and print estimate of the square root of the
! residual sum of squares
If (rank==n) Then
rnorm = dznrm2(m-n, b(n+1), 1)
Write (nout, *)
Write (nout, *) 'Square root of the residual sum of squares'
Write (nout, 110) rnorm
End If
Else
Write (nout, *) 'The SVD algorithm failed to converge'
End If
100 Format (4(' (',F7.4,',',F7.4,')',:))
110 Format (3X, 1P, E11.2)
120 Format (1X, I6)
130 Format (1X, 7F11.4)
End Program
