simrun ❭ modular_reduced_model_inference ❭ reduced_model ❭ Rm ❭ run

Rm.run

simrun.modular_reduced_model_inference.reduced_model.Rm.run(client=None, n_workers=None, strategy_selection=None)

Run one or more strategies on the data.

A strategy is a pipeline that:

  1. Extracts data from the reduced model

  2. Splits the data into training and test sets

  3. Constructs a cost function to be optimized.

  4. Solves the optimization problem using a _Solver

Each strategy implements different cost functions, depending on what to optimize for. However, they all implement a get_score() method to evaluate the performance of the optimization.

Parameters:

  • client (Client) – Dask client for remote optimization

  • n_workers (int) – Amount of workers to use for remote optimization

  • strategy_selection (list) – List of strategy names to run. If None, run all strategies.

See also

simrun.modular_reduced_model_inference.strategy for available strategies.