Re: nonlinear curve fitting using orthogonal distance regression?
- To: mathgroup at smc.vnet.net
- Subject: [mg124731] Re: nonlinear curve fitting using orthogonal distance regression?
- From: Ray Koopman <koopman at sfu.ca>
- Date: Fri, 3 Feb 2012 02:08:26 -0500 (EST)
- Delivered-to: l-mathgroup@mail-archive0.wolfram.com
- References: <jgdmrn$qnv$1@smc.vnet.net>
On Feb 2, 2:00 am, "Michael B. Heaney" <mhea... at gmail.com> wrote:
> Does anyone know if there is a way to do nonlinear curve fitting using
> orthogonal distance regression?
Is this the sort of thing you had in mind?
Dimensions/@{xdata,ydata} -> {{n,p},{n,q}}
f[x_,params_] := function that takes a row of xdata and gives an
estimate of the corresponding row of ydata as a function of params;
FindMinimum[#.#&@Flatten[ f[#,params]&/@xdata - ydata ],
parameter names and initial guesstimates]