Closes #173. Estimates differential rates for background and foreground events, rather that the cumulative rates that were being used previously. The background model is created from a combination of a Gaussian KDE and an exponential fit once there are too few samples to estimate properly. The foreground model is currently just a KDE.
To-do:
[ ] Add tests!
[ ] See whether there's any benefit to optimizing KDE bandwidth.
[ ] Potentially make the foreground model more sophisticated
Closes #173. Estimates differential rates for background and foreground events, rather that the cumulative rates that were being used previously. The background model is created from a combination of a Gaussian KDE and an exponential fit once there are too few samples to estimate properly. The foreground model is currently just a KDE.
To-do: