This is a class structure and notebook implementation for inferring and simulating a type of multi-state models with a counting memory which records previous state occupations. The inference is currently written for pneumococcus-like models but may be generalised to others in the future.

The nifty thing that pneumoinfer does is it implements an approximate ODE description of the system (which would otherwise have to be explicitly simulated) and hence there is a radical reduction in time required to compute the likelihood of the model with respect to a given data set in comparison to other methods. The details behind this mathematical trickery are given in the technical article posted here: umbralcalc.github.io/posts/pneumoinfer.html.