alle_meije
02-16-2006, 05:09 AM
Hi,
I am managing code that I have not written myself, which calls the routines spline() and splins(). The integer n (3rd argument to spline, 4th to splins) is defined as KERNEL = 4.
Could KERNEL, or n, be anything else than 4 for cubic splines? I would guss that it could not be less than 4, but for multidimensional data, it could be as large as dim[i] for which ever dimension you want to interpolate, couoldn't it?
I tried to find references in the book and on the internet, but there are no real explanations as to what values n can take.
Thanks for your help!
I am managing code that I have not written myself, which calls the routines spline() and splins(). The integer n (3rd argument to spline, 4th to splins) is defined as KERNEL = 4.
Could KERNEL, or n, be anything else than 4 for cubic splines? I would guss that it could not be less than 4, but for multidimensional data, it could be as large as dim[i] for which ever dimension you want to interpolate, couoldn't it?
I tried to find references in the book and on the internet, but there are no real explanations as to what values n can take.
Thanks for your help!