Orthogonal distance regression

Please specify the type of your testcase here

The exponential fit: Remember the role of the parameter vector β: β1 is the constant term, next there follow k coefficients of the exponentials and then the k corresponding exponents. Hence the number of the parameters must be odd! Your input below is used as initial guess and possibly for generating the data if you wish so.

a model of your own: Please type the code for computing the function f(x,β) in the textarea below. Remember the role of the first 6 columns! The variables must be named x and beta. You may use some additional variables x1,x2,x3,x4,f1,f2,f3,f4,f5,f6,h,sum, a vector z(100) (index running from 1), the integers i,j,k , logicals bool1,bool2,bool3 and the constants pi(=π), e1(=exp(1)) and sqrt2(=square root(2)). No other variables can be used. As labels you must not use 1 and any number above 89999. c f=x*beta(1)+beta(2)

Specify the number of parameters np in {1,..,10} here. Important: for the exponential model the number must be odd!
np =
Specify the number of data points N (in {npar,..,100}) here
N =

should the data be generated or do you want to provide own x-y-data?

	generate data from parameters: xmin = xmax = errx = in [0,1] ! erry = in [0,1] !
	My own data: Write here a list of N x-y-pairs as 2*N numbers, separated by comma or blank. It is not necessary to use a separate line for each pair. x-y-data = -1.1,-4.2, 0.1,-0.9, 1.1,2.2, 2.05,4.98, 3.1,8.2, 4.05,11.1, 4.99,14.3, 6.1,16.8, 6.99,20.1, 8.06,22.8,

Please specify the parameters β here: either for generating the 'true' data or as initial guesses for fitting your own data: npar numbers separated by comma or blank
3.0, 1.0

you may want to scale the δ_i

	no scaling
	scaling with value =

you may want to use ordinary least squares instead (all δ_i=0)

	no
	yes

You may want to change the standard termination criteria here

	use standard values
	Use these values: sstol = must be in [1.0e-14,1.0e-2] partol = must be in [1.0e-10,1.0e-2] maxiter = must be in [10,10000]

Click on "evaluate", in order to submit your input.