neuroinf-ml

A set of machine learning tools coded as exercises to learn the algorithms more thoroughly. Coded as a group in the DTC for Neuroinformatics at the University of Edinburgh.

Discrete Latent State Markov Model (aka HMM)

Working on this through the burglar example in Barber's textbook. Have so far implemented filtering. This can be run using the burglar.py script, but so far the result isn't plotted (Gavin: wanted to make it write an animated gif). However, it's easy to plot the data however you might like as it's returned by the main() function in burglar.py. So just start up IPython or similar and:

import burglar
filter_results = burglar.main()

For example, if you'd like to plot using the heatmap function we've already written then all you need to do is:

t = 1
import neuroml.plotting
neuroml.plotting.heatmap(filter_results[t])

Where t is the time of the hidden state you want to plot.