Skip to content

SR Approach

SRToolkit.approaches.sr_approach

This module contains the SR_approach class, which is the base class for all symbolic regression approaches.

SR_approach 

SR_approach(name: str)

The base class for all symbolic regression approaches. Any symbolic regression approach should inherit from this class.

Parameters:

Name Type Description Default
name str

The name of the approach.

 required
Source code in SRToolkit/approaches/sr_approach.py 
10
11
12
13
14
15
16
17
18
def __init__(self, name: str):
    """
    The base class for all symbolic regression approaches. Any symbolic regression approach should inherit from
    this class.

    Args:
        name: The name of the approach.
    """
    self.name = name

search 

search(sr_evaluator: SR_evaluator, seed: Optional[int] = None)

Run the symbolic regression search.

Implementations should use the provided evaluator to score generated expressions. All evaluation results are stored inside sr_evaluator, so nothing is returned.

Parameters:

Name Type Description Default
sr_evaluator SR_evaluator

Evaluator used for scoring expressions.

 required
seed Optional[int]

Random seed used for generating expressions.

 None
Source code in SRToolkit/approaches/sr_approach.py 
20
21
22
23
24
25
26
27
28
29
30
31
def search(self, sr_evaluator: SR_evaluator, seed: Optional[int] = None):
    """
    Run the symbolic regression search.

    Implementations should use the provided evaluator to score generated expressions.
    All evaluation results are stored inside `sr_evaluator`, so nothing is returned.

    Args:
        sr_evaluator: Evaluator used for scoring expressions.
        seed: Random seed used for generating expressions.
    """
    raise NotImplementedError

clone 

clone() -> SR_approach

Clones the SR_approach instance. This is used to make multiple independent copies of the approach and making multiple independent evaluations/parallel evaluations of the approach possible. If the approach is stateless, returning self is sufficient, otherwise a deep copy of the approach should be returned. This allows us to do pretraining and finetuning of the approach independently.

Source code in SRToolkit/approaches/sr_approach.py 
33
34
35
36
37
38
39
40
def clone(self) -> "SR_approach":
    """
    Clones the SR_approach instance. This is used to make multiple independent copies of the approach and making
    multiple independent evaluations/parallel evaluations of the approach possible. If the approach is stateless,
    returning self is sufficient, otherwise a deep copy of the approach should be returned. This allows us to do
    pretraining and finetuning of the approach independently.
    """
    raise NotImplementedError