Function docstring

physo.SR(X, y, y_weights=1.0, X_names=None, X_units=None, y_name=None, y_units=None, fixed_consts=None, fixed_consts_units=None, free_consts_names=None, free_consts_units=None, free_consts_init_val=None, op_names=None, use_protected_ops=True, stop_reward=1.0, max_n_evaluations=None, stop_after_n_epochs=10, epochs=None, candidate_wrapper=None, run_config=None, get_run_logger=None, get_run_visualiser=None, parallel_mode=True, n_cpus=None, device='cpu')

Runs a symbolic regression task. (Wrapper around physo.task.fit)

Parameters:

  • X (numpy.array of shape (n_dim, ?,) of float) – Values of the input variables of the problem with n_dim = nb of input variables.

  • y (numpy.array of shape (?,) of float) – Values of the target symbolic function to recover when applied on input variables contained in X.

  • y_weights (np.array of shape (?,) of float or float, optional) – Weight values to apply to y data. Or single float to apply to all (for default value = 1.).

  • X_names (array_like of shape (n_dim,) of str or None (optional)) – Names of input variables (for display purposes).

  • X_units (array_like of shape (n_dim, n_units) of float or None (optional)) – Units vector for each input variables (n_units <= 7). By default, assumes dimensionless.

  • y_name (str or None (optional)) – Name of the root variable (for display purposes).

  • y_units (array_like of shape (n_units) of float or None (optional)) – Units vector for the root variable (n_units <= 7). By default, assumes dimensionless.

  • fixed_consts (array_like of shape (?,) of float or None (optional)) – Values of choosable fixed constants. By default, no fixed constants are used.

  • fixed_consts_units (array_like of shape (?, n_units) of float or None (optional)) – Units vector for each fixed constant (n_units <= 7). By default, assumes dimensionless.

  • free_consts_names (array_like of shape (?,) of str or None (optional)) – Names of free constants (for display purposes).

  • free_consts_units (array_like of shape (?, n_units) of float or None (optional)) – Units vector for each free constant (n_units <= 7). By default, assumes dimensionless.

  • free_consts_init_val (dict of { str : float } or None (optional)) – Dictionary containing free constants names as keys (eg. ‘c0’, ‘c1’, ‘c2’) and corresponding float initial values to use during optimization process (eg. 1., 1., 1.). None will result in the usage of token.DEFAULT_FREE_CONST_INIT_VAL as initial values. None by default.

  • op_names (array_like of shape (?) of str or None (optional)) – Names of choosable symbolic operations (see physo.physym.functions for a list of available operations). By default, uses operations listed in physo.task.args_handler.default_op_names.

  • use_protected_ops (bool (optional)) – If True, uses protected operations (e.g. division by zero is avoided). True by default. (see physo.physym.functions for a list of available protected operations).

  • stop_reward (float (optional)) – Early stops if stop_reward is reached by a program (= 1 by default), use stop_reward = (1-1e-5) when using free constants.

  • max_n_evaluations (int or None (optional)) – Maximum number of unique expression evaluations allowed (for benchmarking purposes). Immediately terminates the symbolic regression task if the limit is about to be reached. The parameter max_n_evaluations is distinct from batch_size * n_epochs because batch_size * n_epochs sets the number of expressions generated but a lot of these are not evaluated because they have inconsistent units.

  • stop_after_n_epochs (int or None (optional)) – Number of additional epochs to do after early stop condition is reached. Uses args_handler.default_stop_after_n_epochs by default.

  • epochs (int or None (optional)) – Number of epochs to perform. By default, uses the number in the default config file.

  • candidate_wrapper (callable or None (optional)) – Wrapper to apply to candidate program’s output, candidate_wrapper taking func, X as arguments where func is a candidate program callable (taking X as arg). By default = None, no wrapper is applied (identity).

  • run_config (dict or None (optional)) – Run configuration (by default uses physo.task.sr.default_config) See physo/config/ for examples of run configurations.

  • get_run_logger (callable returning physo.learn.monitoring.RunLogger or None (optional)) – Run logger (by default uses physo.task.args_handler.get_default_run_logger)

  • get_run_visualiser (callable returning physo.learn.monitoring.RunVisualiser or None (optional)) – Run visualiser (by default uses physo.task.args_handler.get_default_run_visualiser)

  • parallel_mode (bool (optional)) – Parallel execution if True, execution in a loop else. True by default. Overrides parameter in run_config.

  • n_cpus (int or None (optional)) – Number of CPUs to use when running in parallel mode. Uses max nb. of CPUs by default. Overrides parameter in run_config.

  • device (str (optional)) – Device to use for computations (eg. ‘cpu’, ‘cuda’). ‘cpu’ by default.

Returns:

best_expression, run_logger – Best analytical expression found and run logger.

Return type:

physo.physym.program.Program, physo.learn.monitoring.RunLogger