m8m
November 7th, 2005, 05:42 PM
The problem is this:
1) I've got a discrete set of data obtained from an experiment.
2) This set is a convolution of two functions - Gauss and a 2exp decay (zeros at first, then hops to 1 and decays further according to A*exp(-t/a)+B*exp(-t/b) )
3) Nothing else is known
What I need to do is to find parameters of Gauss and 2exp, so that simulated data (convolution of Gauss and 2exp with with these parameters) is approximately the same as obtained. To be more specific, I was told to find a, b and sigma 8))
Any ideas how to do it?
I know how to perform a discrete convolution between two data sets, but I couldn't find an algorithm of deconvolution. So, another question is whether I should simulate the signal (facing that all parameters are unknown) and compare it with obtained, or the other way around - deconvolute obtained data with Gauss and then fit the result with 2exp decay?
1) I've got a discrete set of data obtained from an experiment.
2) This set is a convolution of two functions - Gauss and a 2exp decay (zeros at first, then hops to 1 and decays further according to A*exp(-t/a)+B*exp(-t/b) )
3) Nothing else is known
What I need to do is to find parameters of Gauss and 2exp, so that simulated data (convolution of Gauss and 2exp with with these parameters) is approximately the same as obtained. To be more specific, I was told to find a, b and sigma 8))
Any ideas how to do it?
I know how to perform a discrete convolution between two data sets, but I couldn't find an algorithm of deconvolution. So, another question is whether I should simulate the signal (facing that all parameters are unknown) and compare it with obtained, or the other way around - deconvolute obtained data with Gauss and then fit the result with 2exp decay?