Cubic Spline and its gradient

Hello knowers of things,

I have a set of x and y data points and I'm am trying to get a fit. so far I have followed this process.

x = [...]; %too much data to type here

y = [...];

cs = spline(x,y);

plot(x,y,'o',x,ppval(cs,x),'-'); %this showed an excellent fit

I am using these data points to machine a part but I need the numerical gradient of that spline so can find the magnitude of the normal vector and multiply that by the radius of the machine part and get a new set of x and y coordinates t plug into the mill. When I try the gradient command I get he following output

gradient(cs)

'Error using zeros'

'Trailing string input must be valid numeric class name'

Error in gradient (line 64)

g = zeros(size(f),class(f)); %case of singleton dimension

If anyone could help me out or point me in the right direction I would really appreciate it. I'm still new at this.

[dpb]

Depends on what you need, specifically...

Look at

doc ppval % and friends

to evaluate the spline at a set of points from which one can then do numeric gradient if numeric is sufficient.

Alternatively,

doc unmkpp

will break out the details of the fitted spline from which you can obtain explicit coeffiecients and breakpoints of the piecewise polynomial and do the direct computation of the derivative from the polynomial rule for derivatives.

Answer 1

[Walter Roberson]

Perhaps you should be using fndir() to generate the gradient; you could then evaluate the gradient at particular locations if you want to.

http://www.mathworks.com/support/solutions/en/data/1-15NSB/index.html