MathGroup Archive 2001

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

Search the Archive

Re: Fitting data to line with a specific slope

  • To: mathgroup at smc.vnet.net
  • Subject: [mg30669] Re: [mg30618] Fitting data to line with a specific slope
  • From: Ken Levasseur <Kenneth_Levasseur at uml.edu>
  • Date: Mon, 3 Sep 2001 20:32:45 -0400 (EDT)
  • References: <200108310809.EAA05457@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

Catherine:

You can transform your data first:

In[9]:=
data = {{-1, 5.}, {0, 0.1}, {1, -3}, {2, -6}, {4, -14.3}}

Out[9]=
{{-1, 5.}, {0, 0.1}, {1, -3}, {2, -6}, {4, -14.3}}

m = -3.2;
best =   m x + Fit[data /. {{a_, b_} -> {a, b - m a}}, {1}, x]

Out[19]=
0.2 - 3.2*x

The linearity of least squares fitting lets you do this

Ken Levasseur
Math Sciences
UMass Lowell

Catherine Neish wrote:

>Hi there.
>
>I am attempting to fit my data to a line of the form
>
>    y = -3.2 x + intercept,
>
>but I cannot figure out how to do this with Mathematica.
>
>The function "Fit" takes only basis functions, so I cannot specify that I
>would like the slope to be -3.2.  I also tried "NonlinearFit," but the
>following code
>
>    NonlinearFit[data, intercept - 3.2 x, x, intercept]
>
>yielded the following error message:
>
>Tranpose::nmtx : The first two levels of the one-dimensional list {} cannot
>be transposed
>Tranpose::nmtx : The first two levels of the one-dimensional list {-3.2}
>cannot be transposed
>NonlinearFit::lnnosvd : NonlinearFit was unable to obtain the singular value
>decomposition for the design matrix of the linear model.
>NonlinearFit::fitfail : The fitting algorithm failed.
>
>
>It is possible that my data does not fit a line with slope -3.2 very well.
>Could this be the source of the errors?
>
>Any advice regarding these problems would be greatly appreciated.
>
>Sincerely,
>
>Catherine Neish
>
>
>




  • Prev by Date: Re: convert graphics
  • Next by Date: Diophantine equation y^2=x^3-2*x+1
  • Previous by thread: Re: Fitting data to line with a specific slope
  • Next by thread: Transparent evalaution of a notebook