************80
! L2ERR computes the errors of an L2 approximation.
Discussion:
This routine computes various errors of the current L2 approximation,
whose piecewise polynomial representation is contained in common
block APPROX, to the given data contained in common block DATA.
It prints out the average error ERRL1, the L2 error ERRL2, and the
maximum error ERRMAX.
Modified:
16 February 2007
Author:
Carl DeBoor
Reference:
Carl DeBoor,
A Practical Guide to Splines,
Springer, 2001,
ISBN: 0387953663,
LC: QA1.A647.v27.
Parameters:
Input, integer ( kind = 4 ) IPRFUN. If IPRFUN = 1, the routine prints out
the value of the approximation as well as its error at
every data point.
Output, real ( kind = 8 ) FTAU(NTAU), contains the value of the computed
approximation at each value TAU(1:NTAU).
Output, real ( kind = 8 ) ERROR(NTAU), with
ERROR(I) = SCALE * ( G - F )(TAU(I)). Here, SCALE equals 1
in case IPRFUN /= 1, or the absolute error is greater than 100
somewhere. Otherwise, SCALE is such that the maximum of the
absolute value of ERROR(1:NTAU) lies between 10 and 100. This
makes the printed output more illustrative.
| Type | Intent | Optional | Attributes | Name | ||
|---|---|---|---|---|---|---|
| integer(kind=4) | :: | iprfun | ||||
| real(kind=8) | :: | ftau(ntau) | ||||
| real(kind=8) | :: | error(ntau) |
| Type | Attributes | Name | Initial | |||
|---|---|---|---|---|---|---|
| real | :: | break(lpkmax) | ||||
| real | :: | coef(ltkmax) | ||||
| integer(kind=4) | :: | l | ||||
| integer(kind=4) | :: | k |
| Type | Attributes | Name | Initial | |||
|---|---|---|---|---|---|---|
| integer(kind=4) | :: | ntau |