jsonctmctree overview

The purpose of this project is to provide a relatively easy way to query properties of a continuous-time multivariate finite-state Markov chain on a branching timeline with incomplete observations.

Some of these properties could include the log likelihood, its derivative with respect to logs of edge-specific rate scaling factors, the expected proportions of time spent in subsets of states along certain edges, linear combinations of transition count expectations, and posterior state distributions at nodes.

The properties themselves may be of interest, or they may be only ingredients in a larger algorithm that is out of scope for the jsonctmctree interface, for example point estimation of parameter values.

Until a better user guide is written, the best way to get started is by looking at one of the few examples below.