Fitting data with custom function defined with recursion loop

Hi

I am trying to fit data with a function of this form: the function is defined on strictly positive values. In 0, the function is equal to 1-(the sum of the values it takes for x from 1 to 10)

for example, I tried with the function:

function[b]=Function2(x,a,c)

   if x>0

b=a*(x+c);

   elseif x<0

b=0;

   else

y(1)=0;

for i=2:10

y(i)=y(i-1)+Function2(i,a,c);

end

b=1-y(10);

   end

And used data defined so that I knew what a and c parameters to expect. The curve fitting tool replys to me that "Function value and YDATA sizes are not egal"

When I suppress the loop for x=0 that uses the function Function2 inside it's own expression, the curve fitting tool works (the function is then just a piecewise function), but then it gives a fitting curve that does not correspond at all to the piecewise function I expect.

How can I fit this function?

Thanks in advance

Bill

[Image Analyst]

I've read it 4 times and I still can't understand it. What does "defined on strictly positive values in 0" mean??? And I guess you're talking about element 1-10 of the output b, but when you say "1-(the sum of its positive values, from 1 to 10)", who's positive values are you talking about summing? a's, x's, or c's? And why is Function2 recursive - why is it calling itself?

[Bill Francois]

Hi I've edited my question, it was unclear (mostly for punctuation reasons). What I meant was that the function is defined a certain way for x>0, and for x=0 the function is defined by the sum of some of the values it takes for x>0. c is just a parameter.

[Walter Roberson]

As a debugging step, inside the function please display

size(x), size(a), size(c)

Answer 1

[Titus Edelhofer]

Hi Bill,

I guess Curve Fitting tool will call your function with x being a vector and not a scalar. In this case your function will not work, since it's written for scalar x only. If length(x)>1 you should loop over the vector x.

Titus