Learning functions

[UKMan]

Hello,

Im new to mathcad and im learning everyday from books and the built in help..

i have to find out a few unknowns from the data provided.. I have plotted the graph in the attachment..I have attached the file in pdf to avoid compatibility issues. Ive also attached it in mathcad 13 format... Ive also attached the data sets from which i took the values from

but i dont understand how to use it or what parameters to give it?

Can you please help me out here.. Am i going right till now?

Thank you

[Raiko]

Hello UKman,

I'm not certain what you're after but the attached altered MC sheet may help. It is not an elegant solution but it will help to find peaks.

Basically, what I did is as follows:

• use a fitting function to create a polynomial
  
• extract the coefficients of the fitting function
  

You need to create a polynomial with this coefficients which than can be evaluated as shown in the quicksheet "Finding roots of a polynomial"

Raiko

[UKMan]

Thanks a lot for your replt, Raiko.. but the file you attached is not working..

when in open it in Mathcad 14 it gives me an error saying "An unknown runtime error occurred.. Exiting.."

and after his the program crashes.. can you please look into it..

Thanks a lot for looking into this problem and for helping out..

can you please provide a working file.. Thank you

[MikeArmstrong]

The peak of what curve Hits or ss?

Mike

[UKMan]

Hits is the plot of the actual data.. this is just one example of the data sets I have and most are very noisy.. so thats why i used supsmooth..

so I want to predict or find out the peaks of the ss curve..

Thank you for looking at my post

[MikeArmstrong]

Your file seems to be corrupted.

Try re-saving and re-posting.

Mike

[UKMan]

the file posted by Raiko is corrupt and i requested for a another file.. but the file i attached in the first post.. the mathcad.xmcd is working.. i checked it..

its saved in mathcad 13 version.. can that be a problem? should i save it in mathcad 14?

to avoid these compatibility issues.. i added a pdf printout along with the excel data files..

if the mathcad file doesnt open.. can you please take a look at the pdf I attached.. its a printout of the program.

Thanks a lot.

[MikeArmstrong]

its saved in mathcad 13 version.. can that be a problem? should i save it in mathcad 14?

When possible try and post your worksheets in M11 format for a wider audience.

Mike

[UKMan]

k. thanks.. ill remeber that next time

im not infront of my laptop now.. so i cant get access to mathcad.. ill post a M11 version tomorrow..

but for now can you please look at the pdf to get a general idea.. its just a 3 line program

Thank you

[MikeArmstrong]

I personally don't need M11, but some forum members do.

I am missing something? If you are trying to extract the two highest values of vector ss will the following do?

Mike

[Raiko]

Hi,

sorry for the problems you ran into with that corrupted file.

Anyhow, I've attached a new one with a simpler approach. Just keep in mind that I use a file input for the data which reads directly from the excel file. Unzip the "Voltage" file and right-click on the filoe input icon and choose a new path for data input in the properties dialogue.

Check the MC 13 file whether it helps you

Raiko

[RichardJ]

Don't smooth data that you are going to fit a function to using least squares. Least squares fitting handles noise very well, and smoothing will either have almost no effect or it will bias the parameter estimates. It's better to do the least squares fit using minerr rather than genfit. It's easier to set up and more flexible.

If the heights and/or standard deviations of the two distributions are supposed to be the same you can easily modify the worksheet so that those parameters are scalars rather than two element vectors.

[UKMan]

Thank you everyone for the all the help..

@Richard Jackson

I have sent you a private message. Can you please check it out.

Thank you

[RichardJ]

I should add that your "noise" does not look at all like noise. Noise looks random, and what you have looks like eack peak is really 3 peaks.

[UKMan]

This is an example of a good system. I wanted to work on a good system first before i apply this to very noisy signals.

and that is why i wanted to do smoothing to remove the 3 peaks.. which you correctly pointed out to produce wrong results. Thank you

[RichardJ]

Least squares fitting a function to data is only a valid approach if the function models everything in the data other than the random noise. So if you really have 6 peaks you should fit 6 peaks.

If you have very noisy data where you can't even tell there are more than two peaks then you might run into some problelms. I can't say for sure though without seeing an example of the noisy data.

[UKMan]

I think the most noisy case that I will use will have to be when the two curves overlap slightly.. not have many peaks.. but i dont know yet and will definitly let you know in case of too noisy data giving me problems.