[Date Index] [Thread Index] [Author Index]

Re: surface fitting question

• To: mathgroup at smc.vnet.net
• Subject: [mg61389] Re: surface fitting question
• From: "Ray Koopman" <koopman at sfu.ca>
• Date: Tue, 18 Oct 2005 02:44:30 -0400 (EDT)
• References: <diprth$hpc$1@smc.vnet.net><divgm6$g2n$1@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

Ralph Smith wrote:
> [...]  The Fit[] with polynomials is way off at x=0.
> All of the z values should be 0 but I'm getting
> numbers like 100 or so for the surface along that line.

To force z = 0 at x = 0, try fitting something like

z = a30 x^3 + a21 x^2 y + a12 x y^2 + a20 x^2 + a11 x y + a10 x

which is cubic without the x^0 terms.