pertubation_generator#

Provides the generator for PredictionTask signals.

class PertubationGenerator(sm: SystemModelInterface, species_start_values_noiser: Optional[Noiser] = None, kinetic_parameters_noiser: Optional[Noiser] = None)#

Bases: object

Defines how to generate data for a PertubationTask.

The PertubationGenerator generates data for a PertubationTask by generating a signal for a given system model and check if the signal has a steady state. The initial values for the species and the kinetic parameters are then pertubed. Afterwards the signal is generated again and checked if it has a steady state. If each of the signals has steady states, the data is saved and a table containing the concrete start values for the species, arguments and the according steady-states is returned.

Initializes the PertubationGenerator.

Note

Only the use of a constant_kinetic_parameter is allowed.

Parameters:

  • sm – The system model.

  • species_start_values_noiser – The noiser for the species start values.

  • kinetic_parameters_noiser – The noiser for the kinetic parameters.

generate_signals(n: int = 100) → DataFrame#

Generates signals.

Parameters:

n – The number of samples.

Returns:

A list of (noised and sparsed) signals.

Raises:

ValueError – if a signal has no steady state.

Note

This method will probably not work for prediction tasks using a derivative noiser.