Adding vertex for transition matrices

[johannespetrat]

For models with a time-step (such as Hidden Markov Models) it is often necessary to implement a transition matrix. If we had three states A, B and C and pAB is the probability of transitioning from A to B then we can describe our transition matrix as <a href="https://www.codecogs.com/eqnedit.php?latex=\Bigl(\begin{matrix}&space;p{AA}&space;&&space;p{AB}&p{AC}&space;\&space;p{BA}&space;&&space;p{BB}&p{BC}&space;\&space;p{CA}&space;&&space;p{CB}&p{CC}&space;\end{matrix}\Bigr)&space;\&space;\&space;\text{where&space;}&space;p{AA}&space;+&space;p{AB}&space;+&space;p_{AC}&space;=&space;1" target="blank"><img src="https://latex.codecogs.com/gif.latex?\Bigl(\begin{matrix}&space;p{AA}&space;&&space;p{AB}&p{AC}&space;\&space;p{BA}&space;&&space;p{BB}&p{BC}&space;\&space;p{CA}&space;&&space;p{CB}&p{CC}&space;\end{matrix}\Bigr)&space;\&space;\&space;\text{where&space;}&space;p{AA}&space;+&space;p{AB}&space;+&space;p{AC}&space;=&space;1" title="\Bigl(\begin{matrix} p{AA} & p{AB}&p{AC} \ p{BA} & p{BB}&p{BC} \ p{CA} & p{CB}&p{CC} \end{matrix}\Bigr) \ \ \text{where } p{AA} + p{AB} + p_{AC} = 1" />

This would be really similar to the current implementation of the DirichletVertex. In the DirichletVertex all probabilities sum to 1, but for a transition matrix we only want the rows to sum to 1.

[springcoil]

Plus one on this :)

[sshcherbakov]

Just a quick question. Am I understanding correctly, that state now (Sept. 2020), the Keanu library doesn't support or provide all needed primitives to model Markov chains and process and also cannot be used (alone) to model HMM (Hidden Markov Model)?