| sigma values of polyfit and svdfitr [message #83610] |
Sun, 17 March 2013 23:26  |
gunvicsin11
Messages: 93
Registered: November 2012
|
Member |
|
|
Hello everyone,
Please let me know how to use the sigma values and returned in polyfit and svdfit to find the fitting error. Its very important for my work.
Please help me out in this respect.
thanking you in advance,
sid
|
|
|
|
| Re: sigma values of polyfit and svdfitr [message #83787 is a reply to message #83610] |
Tue, 02 April 2013 15:18  |
Phillip Bitzer
Messages: 223
Registered: June 2006
|
Senior Member |
|
|
The error in the parameters is not exactly the same as the "fitting" error. If you mean "how well does the model fit the data", there are several ways to measure this. One of the more common ways is to use the chi squared statistic. It's one of the output keywords to poly_fit. Chi squared is related to the difference between the input data and the modeled data.
There are a lot of places the assessment of how well a model fits a data set is discussed; two books on my shelf that do a good job addressing this is Data Reduction and Error Analysis by Bevington and Statistical Methods in the Atmospheric Sciences by WIlks.
|
|
|
|
| Re: sigma values of polyfit and svdfitr [message #83798 is a reply to message #83610] |
Mon, 01 April 2013 21:58 |
gunvicsin11
Messages: 93
Registered: November 2012
|
Member |
|
|
On Monday, March 18, 2013 11:56:09 AM UTC+5:30, gunvi...@gmail.com wrote:
> Hello everyone,
>
> Please let me know how to use the sigma values and returned in polyfit and svdfit to find the fitting error. Its very important for my work.
>
> Please help me out in this respect.
>
> thanking you in advance,
>
> sid
Hello sir,
Fitting error in the sense error caused only due to fitting and no other errors are considered, like instrumental error.
My problem is I get sigma values(it is the uncertainty in the polynomial coefficients) very high. So I dont know how to use it to get the fitting error. So I thought I would take (the observed value - fitted value) that will give the fitting error.
Am I right in doing this?
thanking you
sid
|
|
|
|
| Re: sigma values of polyfit and svdfitr [message #83801 is a reply to message #83610] |
Mon, 01 April 2013 08:53 |
Andy Sayer
Messages: 127
Registered: February 2009
|
Senior Member |
|
|
Going back to my first reply: what exactly do you mean by 'fitting error'? If you mean the difference between observations and fitted values, then yes, you will find it by looking at the difference between observations (input data) and fitted values (returned by yfit).
AMS
On Monday, April 1, 2013 1:22:47 AM UTC-4, gunvi...@gmail.com wrote:
> On Monday, March 18, 2013 11:56:09 AM UTC+5:30, gunvi...@gmail.com wrote:
>
>> Hello everyone,
>
>>
>
>> Please let me know how to use the sigma values and returned in polyfit and svdfit to find the fitting error. Its very important for my work.
>
>>
>
>> Please help me out in this respect.
>
>>
>
>> thanking you in advance,
>
>>
>
>> sid
>
>
>
> Hello sir,
>
> So if suppose I am fitting a guassian for a emission profile the (maximum value - fitted maximum) will give the fitting error.
>
> am I correct.
>
> Please let me know.
>
> thanking you
>
> sid
|
|
|
|
| Re: sigma values of polyfit and svdfitr [message #83802 is a reply to message #83610] |
Sun, 31 March 2013 22:22 |
gunvicsin11
Messages: 93
Registered: November 2012
|
Member |
|
|
On Monday, March 18, 2013 11:56:09 AM UTC+5:30, gunvi...@gmail.com wrote:
> Hello everyone,
>
> Please let me know how to use the sigma values and returned in polyfit and svdfit to find the fitting error. Its very important for my work.
>
> Please help me out in this respect.
>
> thanking you in advance,
>
> sid
Hello sir,
So if suppose I am fitting a guassian for a emission profile the (maximum value - fitted maximum) will give the fitting error.
am I correct.
Please let me know.
thanking you
sid
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
