difference using another software package curve fitting

Hello

I have a software package (Kaleidagraph) I've used for curvefitting. This package uses the lagrand-marquandt algoritme to fit curves and an iterative process to go from the initial starting values towards the currently best fits.

When I try to do the same in matlab opts = optimoptions(@lsqcurvefit,'Algorithm','levenberg-marquardt'); [personalParams,personalRes,personalResidual] = lsqcurvefit(heightModel,initialValues,personalData(:,1),personalData(:,2));

I'm fitting formula's containing around 8 parameters. Where the startingvalues for the kaleidagraph software stay nicely in range, the matlab startingvalues often become extreme values that are not representable.

Can anybody point out what could possibly be wrong?

greetings

Niels