DiscreteRegTest#

class DiscreteRegTest(initial_conditions: ndarray, niters: int, parameters: ndarray = None, inputs: ndarray = None, bound: float = None)[source]#

Test case for regularization selection with fully discrete models.

The test_cases argument of opinf.roms.ROM.fit_regselect_discrete() is a list of these.

Parameters:
initial_conditions(n,) ndarray

Initial conditions to be tested.

nitersint

Number of iterations to step forward from the initial conditions.

parameters(p,) ndarray or float or None

Parameter value to be tested.

inputs(m, niters) ndarray or None

Inputs to use in the forward prediction, if the model takes inputs.

boundfloat or None

Amount that the forward prediction is allowed to deviate from the initial conditions without the trajectory being classified as unstable.

Properties:
bound: float = None#
inputs: ndarray = None#
parameters: ndarray = None#
initial_conditions: ndarray#
niters: int#
predict_args: tuple#
Methods:

copy

Return a copy of this test case with new initial conditions.

evaluate

Evaluate the model under the test case conditions.

unstable

Return True if the trajectory Q is unstable.