Eric Denovellis

Speed/Memory Bottlenecks

5

+ Computing the cross-spectra + Minimum phase-decomposition

Some options: + use efficient c code like xarray does with [`rolling`](http://xarray.pydata.org/en/stable/computation.html#rolling-window-operations) and [bottleneck](https://github.com/kwgoodman/bottleneck/tree/master/bottleneck) + continue to use numpy strides, but this doesn't scale well. See [here](http://scikit-image.org/docs/0.10.x/api/skimage.util.html#view-as-windows) and [here](https://stackoverflow.com/questions/4936620/using-strides-for-an-efficient-moving-average-filter)

Makes it more generic for future methods

When there is missing data, the likelihood should be 1 for all states. Add an argument to predict to specify missing data timepoints.

Want the flexibility to add in a local state which directly encodes the animal's position. This requires evaluation of the likelihood at the predict state and passing in position. See...

Unsure if diffusion implementation is correctly handling the boundaries. Need to double check.

Currently, we ignore multiple spikes per bin. This seems to work okay, but we are losing data.