Fitting data with 'gaps' in

[SDmanc]

Hello all,

This may sound a very simple problem but I've been stuck for a short while. Basically I have a data set, see attached, and I want to cut out the peak and fit a polynomial to whats left.

I can see two ways to cut out the middle, by either replacing part of the array with NaNs (non-plottable) or simply deleting the offending section.

Neither seem to produce a new array which lend itself to fitting, I've attached my VI which can do both (needs a rewire to use either for fitting..).

Any ideas?

Regards,

David.

[altenbach]

I am posting via phone, so I cannot look at your data. All fitting vi's take xy data. So all you need is to create a new xy data set, containing the retained indices as x and the retained data as y.

[aeastet]

I think I got what you were looking for. Look at the example that I added and see if this does what you want.

[SDmanc]

Hi,

Thanks for your quick replies. However, aeastet, your VI isn't quite what I'm after.

I know its slightly annoying when people use screen grabs, but if you have a quick look at the one attched - from Origin 7- there I can use a data set with missing y-values to fit. This is what I want to do with my data after I have removed the peak - I'm pretty sure just deleting that part of the array is not the way to go so I've edited my VI accordingly.

After fitting I intend to use the fit as a background to get an absorbtion spectrum, then do further fitting (which I can do okay with another VI - I think!).

Is this any clearer?

Many thanks again.

D.

[altenbach]

Use "polynomial evaluation" using the found coefficients to generate the data for any x-values you possibly want. 😄

[altenbach · Accepted solution]

---

@altenbach wrote:

Use "polynomial evaluation" using the found coefficients to generate the data for any x-values you possibly want. 😄

---

Try this.... (LV 2010)

[SDmanc]

That is, exactly, what i was after. Many, many thanks.

[altenbach]

Here's a much better way. Simply set the weight of the undesired points to zero. 😄

(Modify as needed, there are many things still to improve.)