Dear Buruno, I see the situation. Suppose the number of the sub-interbal is n-1 ( that is [x1.x2],[x2,x3],...,[xn-1,xn]). The number of unknown variables of sub-cubic polynomials is 4n-4. When "spline_type" is "not_a_knot","clamped"or "periodic" ,we have 4n-4 constraints as follows. 2n-2 for data matching and continuity. n-2 for the first derivatives and comtinuity at x2,...,xn-1. n-2 for the second derivatives and continuity at x2,...,xn-1. 2 for the derivative values at near the end points which the above "spline_type" gives.
When "spline_type" is "monotone", the constaraits are as follows. 2n-2 for data mathing and continuity. 2n-2 for the first deribatives mathing and continuity at xi and xi+1 on sub-interval(i=1,...,n-1). These derivatives are calculated by the finite difference method. Best regards. -- View this message in context: http://mailinglists.scilab.org/Monotone-preserving-splin-tp4036235p4036244.html Sent from the Scilab users - Mailing Lists Archives mailing list archive at Nabble.com. _______________________________________________ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
