I'm using Origin Origin 7 with SR2 and found the following bug. If you apply the polynomial function y=a0 + a1x + a2x^2 + ... then the fitted parameter do not correspond to the above ones. Even they are shown as a0, a1, ... they are a1, a2,... and a0 is set to zero! This finding is reported to the support/developers, but the user should know this.
Why I'm using NLSF with polynomial fit? It is to my knowledge the only way, together with setting of constraints of course, to make a fit on two or more separate datasets. This is independend of the finding, but a plausible explanation why nobody obviously found it before. Everybody, and also me, is using the normal polynomial fit from the menu. |
NLSF error with polynomial fit
Posted - 11/12/2002 : 11:30:19 AM
