Help with fitting a exp function

Hi,

I'm trying to fit my data with an exp equation.

Since data gives some noise before the initial (max) value, i was trying to make a filter to cut the data before this max value out with the following code:

function test_amplitude
close all; clc
data = load('enkelt_linie_str.csv');
depth = data(:,1);
amplitude = data(:,2);

k = amplitude==max(amplitude);
filter = depth>depth(k);

amplitude = amplitude(filter)
depth = depth(filter)

% initial values
lambda0 = 1;
A0 = max(amplitude);
p0 = [A0 lambda0];

error = (@(p) sum(power(amplitude-p(1).*exp(-p(2)*depth),2)));
p = fminsearch(error,p0)

A = p(1);
lambda = p(2);

plot(amplitude,depth,'b-'); hold on; axis ij

plot(A*exp(-lambda*depth),depth,'r-')

The fit works without the filter (line 7 to 11) but its a aweful fit and with the filter it gives very small values in return.

I hope you guys can help.

And thanks in advance

/Simon

[Image Analyst]

You forgot to attach your csv file, so probably no one will try anything until you attach it.

[Star Strider]

With simulated data and added noise, it works fine for me. How large are your data sets? How large are they after filter?