I don't have repeated measurements,my application is as follows, I have only one measured spectrum of a part of mouse brain, and that spectrum is suspected to be an addition of several components. I know the spectrum of those components, and as you adviced me, I am trying with multilinear regression, to find the amplitude of each component so that if I add them I get the best fit to the main spectrum. In this way I should have an estimate of concentration or to know if this component is present or not.
do you advice me to continue with this approach, and maybe take several measurements to enhance the statistics?
Thank you very much Tom ,multilinear regression worked fine,
but how to avoid getting negative coefficients? how to add constraints on the multilinear regression to get only coefficients with physical meaning, not just coefficients which results in a good fit.
very useful program, but how can we predefine the position of the fitted peaks. this could be helpful for spectroscopy application where you can fit predetermined peaks or functions in a main plot that contain the contribution of all the sub peaks. in other words when you have a main plot, which has to be fitted using multiple peaks with known characteristics.