| Re: Reducing a set of curves to a mean curve [message #69342] |
Tue, 05 January 2010 03:58 |
Bernhard Reinhardt
Messages: 26
Registered: October 2008
|
Junior Member |
|
|
Bernhard Reinhardt wrote:
> Hi everyone,
>
> I have a set of curves (as x-y-points) which I'd like to reduce to a
> mean curve and some kind of deviation interval. If the curves would
> consist of fixed x-values with varying y-values this would be no big
> deal since I could compute the mean and a deviation measure at every
> x-value. However it turns out that the x-values are not fixed as well.
>
> My first idea was to use poly_fit but the results don't look like I want
> them to and it doesn't solve my problem with the deviation interval.
> Right now I'm thinking of using a fixed x-grid and interpolating the
> data. But before I reinvent the wheel again I want to ask if someone
> here has appropriate code in his library. I think this should be a
> common problem!?
Well, I did now linear interpolation to a fixed x-grid. Easier then I
thought with INTERPOL.
Regards,
Bernhard
|
|
|
|
| Re: Reducing a set of curves to a mean curve [message #69348 is a reply to message #69342] |
Mon, 04 January 2010 13:03  |
Vince Hradil
Messages: 574
Registered: December 1999
|
Senior Member |
|
|
On Jan 4, 9:59 am, Bernhard Reinhardt
<wirdseltengele...@freisingnet.de> wrote:
> Hi everyone,
>
> I have a set of curves (as x-y-points) which I'd like to reduce to a
> mean curve and some kind of deviation interval. If the curves would
> consist of fixed x-values with varying y-values this would be no big
> deal since I could compute the mean and a deviation measure at every
> x-value. However it turns out that the x-values are not fixed as well.
>
> My first idea was to use poly_fit but the results don't look like I want
> them to and it doesn't solve my problem with the deviation interval.
> Right now I'm thinking of using a fixed x-grid and interpolating the
> data. But before I reinvent the wheel again I want to ask if someone
> here has appropriate code in his library. I think this should be a
> common problem!?
>
> Best regrads!
>
> Bernhard
Does this help: http://tinyurl.com/yz6hdx3
|
|
|
|
| Re: Reducing a set of curves to a mean curve [message #69351 is a reply to message #69348] |
Mon, 04 January 2010 11:02 |
rogass
Messages: 200
Registered: April 2008
|
Senior Member |
|
|
On 4 Jan., 16:59, Bernhard Reinhardt
<wirdseltengele...@freisingnet.de> wrote:
> Hi everyone,
>
> I have a set of curves (as x-y-points) which I'd like to reduce to a
> mean curve and some kind of deviation interval. If the curves would
> consist of fixed x-values with varying y-values this would be no big
> deal since I could compute the mean and a deviation measure at every
> x-value. However it turns out that the x-values are not fixed as well.
>
> My first idea was to use poly_fit but the results don't look like I want
> them to and it doesn't solve my problem with the deviation interval.
> Right now I'm thinking of using a fixed x-grid and interpolating the
> data. But before I reinvent the wheel again I want to ask if someone
> here has appropriate code in his library. I think this should be a
> common problem!?
>
> Best regrads!
>
> Bernhard
Hi,
SPLINE_PP or SVDFIT may solve your problems. You might also reform
your x-data into one big data set and also the y-data (take the sort
index from x) to enable a broader set of fitting approaches.
Viel Glueck
CR
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
