help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Curve fitting - polynomial fit help SOLVED


From: Rob Teasdale
Subject: Re: Curve fitting - polynomial fit help SOLVED
Date: Tue, 29 May 2007 16:37:50 -0700 (PDT)

Hi Muthu,



Thanks for your reply. Your suggestion about coefficients led me to discover my 
problem.  I was reading my experimental data from a csv file, and the first 
line of the text was labels etc. This was read by dlmread as a zero and 
consequently my data was considerably affected. All I have to do know is work 
out how to tell dlmread to start reading from the second row (the help says 
that I should be able to call it with a row start, but I get a error: range 
must include [R1 C1 R2 C2]) and I will be set.

Thanks again,
Rob

---- Original Message ----

From: Muthiah Annamalai <address@hidden>

To: Xierxes <address@hidden>

Sent: Saturday, 26 May, 2007 5:06:18 AM

Subject: Re: Curve fitting - polynomial fit help



On 5/25/07, Xierxes <address@hidden> wrote:

>

> Hi all,

>

> I am new to Octave and I am trying to fit some experimental data using a

> simple quadratic. I have already fit the data in gnuplot, but I need to

> develop more complex fitting curves - hence my move to Octave.  The problem

> that I am having is that the coefficients generated by gnuplot do not match

> the ones generated by the polyfit, wpolyfit or leasqr (I modified the

> leasqrdemo to use a quadratic fit, and suit my data) functions. However when

> plotted the gnuplot fit looks much closer to experimental data.  I am new to

> curve fitting and as a result I am not sure where to go next.

>

> I thought that gnuplot and leasqr used the same method for calculating the

> coefficients, hence I am somewhat lost. Anyway back to my probability and

> statistics for the engineer and scientist - Any help anyone could provide

> would be fantastic,

> Cheers

> Rob

>

> I am using the latest Octave (2.9.12 and the Optim package from

> Octave-Forge).

> --

> View this message in context: 
> http://www.nabble.com/Curve-fitting---polynomial-fit-help-tf3818989.html#a10812300

> Sent from the Octave - General mailing list archive at Nabble.com.

>

> _______________________________________________

> Help-octave mailing list

> address@hidden

> https://www.cae.wisc.edu/mailman/listinfo/help-octave

>



Maybe you are using the returned polynomial coefficients from polyfit() in the

converse ordering. Is that of some help?



But again, what 'fit' does GNU Plot make? If it is a non-linear fit,

then I guess

we cannot compare the outputs with Octave. Nonlinear fits are very sensitive

to initial guess parameters; maybe you should play with them more.



Cheers,

Muthu









      
_________________________________________________________________________________
              

How would you spend $50,000 to create a more sustainable environment in 
Australia? Go to Yahoo!7 Answers and share your idea.
http://advision.webevents.yahoo.com/aunz/lifestyle/answers/y7ans-babp_reg.html





reply via email to

[Prev in Thread] Current Thread [Next in Thread]