| Re: xerr [message #64272] |
Wed, 17 December 2008 14:41  |
laxsri
Messages: 6
Registered: November 2008
|
Junior Member |
|
|
On Dec 18, 9:08 am, Paolo <pgri...@gmail.com> wrote:
> This is discussed for example in
> section 15.3 in edition 3 of the book
> "numerical recipes".
>
Not sure how to constrain the intercept though...
It seemed rather easy to use mpfitfun! Wondering if that is wrong?
Lakshmi
> Ciao,
> Paolo
>
> Vince Hradil wrote:
>> On Dec 17, 2:27 pm, lakshmi <lax...@gmail.com> wrote:
>>> Hi,
>
>>> I've been using mpfitfun to fit measured values of period (y) and
>>> distances (x) in a linear equation y = a + bx.
>>> I would like to know if we can include the measured uncertainties in x
>>> values too?
>
>>> Thanks,
>
>>> Lakshmi
>
>> Well, since it's a linear problem you should probably choose a linear
>> solution, not mpfitfun. Also, you need to take into account the
>> variance and covariance for both x and y, so you need to solve this
>> with care.
>
>> If you google "fitting a straight line when both variables are subject
>> to error" you'll get a lot of info:http://tinyurl.com/54m8l3
|
|
|
|
|
|
| Re: xerr [message #64274 is a reply to message #64273] |
Wed, 17 December 2008 14:08  |
pgrigis
Messages: 436
Registered: September 2007
|
Senior Member |
|
|
This is discussed for example in
section 15.3 in edition 3 of the book
"numerical recipes".
Ciao,
Paolo
Vince Hradil wrote:
> On Dec 17, 2:27�pm, lakshmi <lax...@gmail.com> wrote:
>> Hi,
>>
>> I've been using mpfitfun to fit measured values of period (y) and
>> distances (x) in a linear equation y = a �+ bx.
>> I would like to know if we can include the measured uncertainties in x
>> values too?
>>
>> Thanks,
>>
>> Lakshmi
>
> Well, since it's a linear problem you should probably choose a linear
> solution, not mpfitfun. Also, you need to take into account the
> variance and covariance for both x and y, so you need to solve this
> with care.
>
> If you google "fitting a straight line when both variables are subject
> to error" you'll get a lot of info: http://tinyurl.com/54m8l3
|
|
|
|
|
|
| Re: xerr [message #64347 is a reply to message #64274] |
Thu, 18 December 2008 19:24 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
On Dec 17, 5:08 pm, Paolo <pgri...@gmail.com> wrote:
> This is discussed for example in
> section 15.3 in edition 3 of the book
> "numerical recipes".
I've used the Numerical Recipes hack for X errors successfully before.
As mentioned, orthogonal distance regression is the real way to do
this, but unfortunately MPFIT does not support this. [ It could in
principle with a lot of work, but doesn't in practice. ]
Craig
|
|
|
|
| Re: xerr [message #64358 is a reply to message #64274] |
Thu, 18 December 2008 06:49 |
Jeremy Bailin
Messages: 618
Registered: April 2008
|
Senior Member |
|
|
On Dec 17, 5:08 pm, Paolo <pgri...@gmail.com> wrote:
> This is discussed for example in
> section 15.3 in edition 3 of the book
> "numerical recipes".
>
> Ciao,
> Paolo
>
> Vince Hradil wrote:
>> On Dec 17, 2:27 pm, lakshmi <lax...@gmail.com> wrote:
>>> Hi,
>
>>> I've been using mpfitfun to fit measured values of period (y) and
>>> distances (x) in a linear equation y = a + bx.
>>> I would like to know if we can include the measured uncertainties in x
>>> values too?
>
>>> Thanks,
>
>>> Lakshmi
>
>> Well, since it's a linear problem you should probably choose a linear
>> solution, not mpfitfun. Also, you need to take into account the
>> variance and covariance for both x and y, so you need to solve this
>> with care.
>
>> If you google "fitting a straight line when both variables are subject
>> to error" you'll get a lot of info:http://tinyurl.com/54m8l3
>
>
On a complete tangent... how is the third edition compared to the
second? I've been hemming and hawing about picking it up.
-Jeremy.
|
|
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
