Two Python implementations of fitting a sum of exponentials to numerical data. I haven't tested both carefully, but method 2 seems to work better:

1. Based on Kaufmann (2003) "Fitting a Sum of Exponentials to Numerical Data".
2. Based on scipy.optimize differential evolution

Note: The kaufmann implementation finds the best fit assuming the exponential does not decay to zero, whereas the diffevol implementation does assume the exponential decays to zero.

##TODO

• add error analysis (note: proper error analysis would probably include basin hopping, but that's probably too sophosticated)
• describe algorithm