probability transition

Terms from Artificial Intelligence: humans at the heart of algorithms

Page numbers are for draft copy at present; they will be replaced with correct numbers when final book is formatted. Chapter numbers are correct and will not change now.

In a stochastic system the probabilty transition is the probabilty that the state will change (transition) to another state. This is often expressed in a transition table that lists for each 'state to next state' pair the probability of the particuar transition. In the case of a Markov process this is all of the information you need to work out future behaviour, but if the system has soem form of hidden histiry, then a more complex probabilty transition table might be needed.

Used in Chap. 14: page 226

Estimated transition probabilities based on observed transition frequencies

hidden Markov model for British summer weather – hidden state in brackets (Changeable/Wet)