Open filtron opened 11 months ago
It might be more natural to think of a Markov kernel as a family of probability distributions indexed by some parameter $x \in \mathbb{X}$. Thus, to retrieve the probability distribution for a particular $x$ it might be more natural to write
markov_kernel(x)
rather than the current
condition(markov_kernel, x)
It might be more natural to think of a Markov kernel as a family of probability distributions indexed by some parameter $x \in \mathbb{X}$. Thus, to retrieve the probability distribution for a particular $x$ it might be more natural to write
rather than the current