replay_trajectory_classification.discrete_state_transitions.DiagonalDiscrete#

class DiagonalDiscrete(diagonal_value: float = 0.98)[source]#

Bases: object

Transition matrix with diagonal_value on the value for n_states

Off-diagonals are probability: (1 - diagonal_value) / (n_states - 1)

diagonal_value#
Type:

float, optional

Methods

make_state_transition(n_states)

Makes discrete state transition matrix.

Methods

make_state_transition

Makes discrete state transition matrix.

Attributes

diagonal_value

make_state_transition(n_states: int) → ndarray[source]#

Makes discrete state transition matrix.

Parameters:

n_states (int)

Returns:

discrete_state_transition

Return type:

np.ndarray, shape (n_states, n_states)