eckity.sklearn_compatible.classification_evaluator

 1"""
 2This module implements the fitness evaluation class, which delivers the fitness function.
 3You will need to implement such a class to work with your own problem and fitness function.
 4"""
 5import numpy as np
 6from sklearn.metrics import accuracy_score
 7
 8from eckity.evaluators.simple_individual_evaluator import SimpleIndividualEvaluator
 9
10CLASSIFICATION_THRESHOLD = 0
11
12
13class ClassificationEvaluator(SimpleIndividualEvaluator):
14    """
15    Class to compute the fitness of an individual in classification problems.
16    """
17
18    def __init__(self, X=None, y=None, metric=accuracy_score):
19        super().__init__()
20        self.X = X
21        self.y = y
22        self.metric = metric
23
24    def set_context(self, context):
25        """
26        Receive X and y values and assign them to X and y fields.
27
28        Parameters
29        ----------
30        context: tuple. first element is a numpy array of size (n_samples, n_features),
31                        and the second element is a numpy array of size (n_samples, 1) or (n_samples,)
32            X matrix and y vector, either (X_train, y_train) or (X_test, y_test), depending on the evolution stage
33
34        Returns
35        -------
36        None.
37        """
38        self.X = context[0]
39        self.y = context[1]
40
41    def evaluate_individual(self, individual):
42        """
43        Compute the fitness value by comparing the program tree execution result to the result vector y
44
45        Parameters
46        ----------
47        individual: Tree
48            An individual program tree in the GP population, whose fitness needs to be computed.
49            Makes use of GPTree.execute, which runs the program.
50            Calling `GPTree.execute` must use keyword arguments that match the terminal-set variables.
51            For example, if the terminal set includes `x` and `y` then the call is `GPTree.execute(x=..., y=...)`.
52
53        Returns
54        -------
55        float:
56            computed fitness value
57        """
58        y_pred = self.classify_individual(individual)
59        return self.metric(y_true=self.y, y_pred=y_pred)
60
61    def classify_individual(self, individual):
62        return np.where(individual.execute(self.X) > CLASSIFICATION_THRESHOLD, 1, 0)