From: Zahra on
Hi all,

I am calculating the errors for a fit to a data set using:

[Q,R] = qr(jacobian,0);
Rinv = inv(R);
se = sqrt(sum(Rinv.*Rinv,2)*resnorm/dfe);

I later need to import these error bars into a spreadsheet. However, the se elements are of the follwoing form:

(1,1) 1.3053
(2,1) 1.2639

and so on which makes it difficult to import into a spreadsheet. How could I get rid of the (1,1) and (2,1) so I have only an array of the errors bars?

Any help is greatly appreciated! Merci.
From: Walter Roberson on
Zahra wrote:

> se = sqrt(sum(Rinv.*Rinv,2)*resnorm/dfe);

> I later need to import these error bars into a spreadsheet. However, the
> se elements are of the follwoing form:
> (1,1) 1.3053
> (2,1) 1.2639
> and so on which makes it difficult to import into a spreadsheet. How
> could I get rid of the (1,1) and (2,1) so I have only an array of the
> errors bars?

How are you examining the se elements? Are you copying from the Variable
Editor window perhaps? Or is your jacobian matrix possibly sparse, resulting
in a sparse calculation throughout ??

> Rinv = inv(R);

There is very likely a better solution than using inv().
From: Zahra on
No matter how I examine the se elements, I always get the associated indices. Either by looking at them in the the Variable Editor window, or if for example I use:

A=se(1,1), it will just write the associated indices (1,1) for A as well...

Thank you,



Walter Roberson <roberson(a)hushmail.com> wrote in message <i3aceb$30m$1(a)canopus.cc.umanitoba.ca>...
> Zahra wrote:
>
> > se = sqrt(sum(Rinv.*Rinv,2)*resnorm/dfe);
>
> > I later need to import these error bars into a spreadsheet. However, the
> > se elements are of the follwoing form:
> > (1,1) 1.3053
> > (2,1) 1.2639
> > and so on which makes it difficult to import into a spreadsheet. How
> > could I get rid of the (1,1) and (2,1) so I have only an array of the
> > errors bars?
>
> How are you examining the se elements? Are you copying from the Variable
> Editor window perhaps? Or is your jacobian matrix possibly sparse, resulting
> in a sparse calculation throughout ??
>
> > Rinv = inv(R);
>
> There is very likely a better solution than using inv().
From: Jan Simon on
Dear Zahra,

> I later need to import these error bars into a spreadsheet. However, the se elements are of the follwoing form:
>
> (1,1) 1.3053
> (2,1) 1.2639

The variable "se" is sparse. You can convert it to a full array:
full(se)

Kind regards, Jan
From: Zahra on
Thank you so much Jan, full(se) works perfectly, I now have all of the errorbars in full arrya format. Thanks again!


"Jan Simon" <matlab.THIS_YEAR(a)nMINUSsimon.de> wrote in message <i3ad8e$17$1(a)fred.mathworks.com>...
> Dear Zahra,
>
> > I later need to import these error bars into a spreadsheet. However, the se elements are of the follwoing form:
> >
> > (1,1) 1.3053
> > (2,1) 1.2639
>
> The variable "se" is sparse. You can convert it to a full array:
> full(se)
>
> Kind regards, Jan