Datafitting - Cubic Spline Boundary Conditions

[Page__Chris]

I would like to fit a cubic spline with the second derivatives at the end points linearly extrapolated from the two closest interior points. Is there a way to specify this type of boundary condition?

[Pavel_D_Intel1]

Hi Chris,

Intel MKL Data Fitting supports next types of boundary conditions for cubic splines:

• not-a-knot
• free end
• periodic
• user-definedе the first/second derivative at the left/right endpoint

You can set the specifically defined second derivatives at the endpoints as presented below:

…
MKL_INT bc_type;             // boundary conditions type
double bc[2];                      // array of boundary conditions
…
bc_type     = DF_BC_2ND_LEFT_DER | DF_BC_2ND_RIGHT_DER;
bc[0] = a;
bc[1] = b;
…
// Call dfdEditPPSpline1D( task, sorder, stype, bc_type, bc, ic_type, ic, scoeff, scoeffhint );
// Call dfdConstruct1D( task, DF_PP_SPLINE, DF_METHOD_STD );

Full code can be found in “dfdcubicsplineint2der.c” sample.

Feel free to ask more.

Best regards,
Pavel.