In a time series or other form of sequential data, the transition probability measures the conditional probability of moving from one state to another. For example, if it is sunny today, we may know from historical data that the probabilty it will also be sunny tomorrow is 0.4. Tranistion probabilities can also be calculated based on more than one past state, for example, if the last two days were dry, how likely is it to also be dry tomorrow. Transition probabilities can be used to create a Markov model of the process.
Used in Chap. 14: pages 208, 209