LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Linear combination analysis

Hi all:
 
I need to do a linear combination of a spectra using two or more standards, I wonder if anyone has done this before or is there a simple way to do it. I am new in Labview.
 
Many thanks
0 Kudos
Message 1 of 7
(3,888 Views)
Hi there,
 
Can you elaborate in a little more detail on what you are trying to do?
 
Regards,
Way S.
NI UK Applications Engineer
0 Kudos
Message 2 of 7
(3,850 Views)
Hi Way:
 
I managed to do the linear combination analysis but still need some help in relation to grouping similar data.
 
I have 100 X-ray spectra and I have taken two of them as standard i.e. one spectra with lots of peaks (standard1) and another spectra with no peaks (standard2). I have done the linear combination of these two standards to produce each of the 100 spectra and as a result I have got the percentage of standard 1 and stardard 2 in each of the 100 spectra.
 
Now I need a statistical method in which I can group similar spectra, for that I need a better way to compare two spectra than the difference between them. Is there any VI out there that can compare spectra i.e. include shifts in the spectra and takes into account different peak amplitudes?. 
 
Many thanks
 
Amb
0 Kudos
Message 3 of 7
(3,804 Views)
I have a difficult time imagining your exact problem. Can you show us an image with a few of the spectra.
 
For the linear combination, I am assuming you are using the "general LS linear fit".
 
If the peak amplitudes vary, you just need a bigger set of basis spectra, one for each unique peak. Shifts are a bit more difficult. Are individual peaks shifted or are all shifted together, e.g. because of an uncalibrated x-axis?
 
Do you have a mathematical model for the spectra? Maybe it can be fit from first principles using e.g. Levenberg-Marqardt?
0 Kudos
Message 4 of 7
(3,794 Views)

Hi,

Thanks for your replies. I know the name of the method I need to use and is called: "Cluster analysis" to build a "Hierarchical Tree". Do you know if there is any VI outthere that can do this?

Cheers

Angela

 

 

 

 

 

0 Kudos
Message 5 of 7
(3,768 Views)
Hi Angela,
Hope all is good. I am having a hard time trying to understand what you mean by "Cluster analysis" to build a "Hierarchical Tree". Could you elaborate your question with perhaps an explanation or example of what you are after so I can point you to the right direction?
Thank you.
0 Kudos
Message 6 of 7
(3,712 Views)

Hi all:

 

Thanks for you response. I am trying to develop soemthing like the program Polysnap http://www.chem.gla.ac.uk/snap/ in Labview. I am attaching a figure of the cluster analysis. Each of the rectangles is one spectra, the ones that are very similar are grouped in clusters. This is called clustering analysis: http://www.statsoft.com/textbook/stathome.html?stcluan.html&1.

Can I do this in Labview?

Many thanks

 

amb

0 Kudos
Message 7 of 7
(3,669 Views)