matlab svd double,fortran lapack的dgesvd的结果和matlab的SVD分解做出来的结果不一致，请大神指点！ - 程序语言 - 小木虫 - 学术 科研 互动社区...

对于A矩阵，matlab SVD分解结果如下：

A =

2.2700   -1.5400    1.1500   -1.9400

0.2800   -1.6700    0.9400   -0.7800

-0.4800   -3.0900    0.9900   -0.2100

1.0700    1.2200    0.7900    0.6300

-2.3500    2.9300   -1.4500    2.3000

0.6200   -7.3900    1.0300   -2.5700

>> [U,S,V] = svd(A,0)

U =

0.2774   -0.6003   -0.1277    0.1323

0.2020   -0.0301    0.2805    0.7034

0.2918    0.3348    0.6453    0.1906

-0.0938   -0.3699    0.6781   -0.5399

-0.4213    0.5266    0.0413   -0.0575

0.7816    0.3353   -0.1645   -0.3957

S =

9.9966         0         0         0

0    3.6831         0         0

0         0    1.3569         0

0         0         0    0.5000

V =

0.1921   -0.8030    0.0041   -0.5642

-0.8794   -0.3926   -0.0752    0.2587

0.2140   -0.2980    0.7827    0.5027

-0.3795    0.3351    0.6178   -0.6017

但在linux用fortran lapack得到的结果如下：

主函数：

*     DGESVD Example Program Text

*     NAG Copyright 2005.

*     .. Parameters ..

INTEGER          NIN, NOUT

PARAMETER        (NIN=5,NOUT=6)

INTEGER          MMAX, NB, NMAX, IERR

PARAMETER        (MMAX=10,NB=64,NMAX=8)

INTEGER          LDA, LDVT, LWORK

PARAMETER        (LDA=MMAX,LDVT=NMAX,

+                 LWORK=MMAX+4*NMAX+NB*(MMAX+NMAX))

*     .. Local Scalars ..

DOUBLE PRECISION EPS, SERRBD

INTEGER          I, IFAIL, INFO, J, LWKOPT, M, N

*     .. Local Arrays ..

DOUBLE PRECISION A(LDA,NMAX), DUMMY(1,1), RCONDU(NMAX),

+                 RCONDV(NMAX), S(NMAX), UERRBD(NMAX),

+                 VERRBD(NMAX), VT(LDVT,NMAX), WORK(LWORK)

*     .. External Functions ..

DOUBLE PRECISION DLAMCH

EXTERNAL         DLAMCH

*     .. External Subroutines ..

EXTERNAL         DDISNA, DGESVD, X04CAF

*     .. Executable Statements ..

OPEN(NIN,FILE='dgesvd-ex.d',STATUS='OLD',IOSTAT=IERR)

OPEN(NOUT,FILE='dgesvd-ex.r',STATUS='REPLACE',IOSTAT=IERR)

WRITE (NOUT,*) 'DGESVD Example Program Results'

WRITE (NOUT,*)

*     Skip heading in data file

READ (NIN,*)

READ (NIN,*) M, N

IF (M.LE.MMAX .AND. N.LE.NMAX) THEN

*

*        Read the m by n matrix A from data file

*

READ (NIN,*) ((A(I,J),J=1,N),I=1,M)

*

*        Compute the singular values and left and right singular vectors

*        of A (A = U*S*(V**T), m.ge.n)

*

CALL DGESVD('Overwrite A by U','Singular vectors (V)',M,N,A,

+               LDA,S,DUMMY,1,VT,LDVT,WORK,LWORK,INFO)

LWKOPT = WORK(1)

*

IF (INFO.EQ.0) THEN

*

*           Print solution

*

WRITE (NOUT,*) 'Singular values'

WRITE (NOUT,99999) (S(J),J=1,N)

*

IFAIL = 0

CALL X04CAF('General',' ',M,N,A,LDA,

+                  'Left singular vectors (first n columns of U)',

+                  IFAIL)

WRITE (NOUT,*)

CALL X04CAF('General',' ',N,N,VT,LDVT,

+                  'Right singular vectors by row (V**T)',IFAIL)

*

*           Get the machine precision, EPS and compute the approximate

*           error bound for the computed singular values.  Note that for

*           the 2-norm, S(1) = norm(A)

*

EPS = DLAMCH('Eps')

SERRBD = EPS*S(1)

*

*           Call DDISNA (F08FLF) to estimate reciprocal condition

*           numbers for the singular vectors

*

CALL DDISNA('Left',M,N,S,RCONDU,INFO)

CALL DDISNA('Right',M,N,S,RCONDV,INFO)

*

*           Compute the error estimates for the singular vectors

*

DO 20 I = 1, N

UERRBD(I) = SERRBD/RCONDU(I)

VERRBD(I) = SERRBD/RCONDV(I)

20       CONTINUE

*

*           Print the approximate error bounds for the singular values

*           and vectors

*

WRITE (NOUT,*)

WRITE (NOUT,*) 'Error estimate for the singular values'

WRITE (NOUT,99998) SERRBD

WRITE (NOUT,*)

WRITE (NOUT,*)

+        'Error estimates for the left singular vectors'

WRITE (NOUT,99998) (UERRBD(I),I=1,N)

WRITE (NOUT,*)

WRITE (NOUT,*)

+        'Error estimates for the right singular vectors'

WRITE (NOUT,99998) (VERRBD(I),I=1,N)

ELSE

WRITE (NOUT,99997) 'Failure in DGESVD. INFO =', INFO

END IF

*

*        Print workspace information

*

IF (LWORK.LT.LWKOPT) THEN

WRITE (NOUT,*)

WRITE (NOUT,99996) 'Optimum workspace required = ', LWKOPT,

+        'Workspace provided         = ', LWORK

END IF

ELSE

WRITE (NOUT,*) 'MMAX and/or NMAX too small'

END IF

STOP

*

99999 FORMAT (3X,(8F8.4))

99998 FORMAT (4X,1P,6E11.1)

99997 FORMAT (1X,A,I4)

99996 FORMAT (1X,A,I5,/1X,A,I5)

END

输入文件dgesvd-ex.d为：

DGESVD Example Program Data

6      4                   :Values of M and N

2.27  -1.54   1.15  -1.94

0.28  -1.67   0.94  -0.78

-0.48  -3.09   0.99  -0.21

1.07   1.22   0.79   0.63

-2.35   2.93  -1.45   2.30

0.62  -7.39   1.03  -2.57  :End of matrix A

输出结果文件dgesvd-ex.r为：

DGESVD Example Program Results

Singular values

9.9966  3.6831  1.3569  0.5000

Left singular vectors (first n columns of U)

1         2            3          4

1  -0.2774 -0.6003 -0.1277  0.1323

2  -0.2020 -0.0301  0.2805  0.7034

3  -0.2918  0.3348  0.6453  0.1906

4   0.0938 -0.3699  0.6781 -0.5399

5   0.4213  0.5266  0.0413 -0.0575

6  -0.7816  0.3353 -0.1645 -0.3957

Right singular vectors by row (V**T)

1          2           3            4

1  -0.1921  0.8794 -0.2140  0.3795

2  -0.8030 -0.3926 -0.2980  0.3351

3   0.0041 -0.0752  0.7827  0.6178

4  -0.5642  0.2587  0.5027 -0.6017

Error estimate for the singular values

1.1E-15

Error estimates for the left singular vectors

1.8E-16    4.8E-16    1.3E-15    2.2E-15

Error estimates for the right singular vectors

1.8E-16    4.8E-16    1.3E-15    1.3E-15

这里输出的是V的转置，为什么U、V都是第一列符号相反呢？请大神指教！非常感谢