implicit fitting algorithm
hi agan igor all
another attribution question. the fitting with user generated functions, fitfunc has the option to fit data without specifying any independent variables. it is very handy for closed curve functions in 2d where there are multiple y values for every x, and cases where the function can be easy arranged to have a zero, but the variables can't be conveniently written to isolate a dependent.. there was a discussion in the late teens on this forum about fitting an ellipse, where i became aware of this handy capability
besides having the ability to do this routinely, a wonderful gift from our heroes at wm, it works extremely well given a decent first guess. astonishingly well, often in a single iteration. there is not a lot of documentation on this and no reference to the source algorithm. is there a citation or is this another wavemetrics innovation? nick
Hello, Nick! Still can't find the shift key, I see 😏
DisplayHelpTopic "Fitting Implicit Functions"
DisplayHelpTopic "Curve Fitting References"
In the references brought up by the second command there are some references under "References for the ODRPACK95 package used for orthogonal distance regression:" Those apply also to implicit functions.
March 5, 2026 at 10:00 am - Permalink
hi john
thanks! yes, still too lazy to press the shift key is the way it is usually expressed. ah ha! that's some good documentation. i'm set. n
March 6, 2026 at 04:56 pm - Permalink