DiscreteTimeStateTransitionModel.transition_prob_matrix

DiscreteTimeStateTransitionModel.transition_prob_matrix(events)

Compute the Markov transition probability matrix.

Parameters:

events (None | ArrayLike) – None, or an array of shape (T, N, R), where T is the number of timesteps, N=self.num_units` the number of units and ``R=self.incidence_matrix.shape[0] the number of transitions.

Returns:

Transition probabilty matrix. If events is None, this matrix is of shape (N, S, S), transition probability matrix associated with the initial state (self.initial_state), of shape (N, S), where S=self.num_states is the number of distinct states. Otherwise, this matrix is of shape (T, N, S, S), representing the transition probability matrix at each timestep. of shape (T, N, S, S).

Return type:

Array