def mock_ale_explainer(request):
"""
Instantiates an ALE explainer with a mock predictor.
"""
out_dim, out_type = request.param
predictor = MockPredictor(out_dim=out_dim, out_type=out_type, seed=0)
explainer = ALE(predictor)
return explainer
def make_ale(dirname: Optional[Path] = None) -> ALE:
data = load_iris()
feature_names = data.feature_names
target_names = data.target_names
X = data.data
y = data.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.25,
random_state=42)
# train model
lr = LogisticRegression(max_iter=200)
lr.fit(X_train, y_train)
# create explainer
logit_fun_lr = lr.decision_function
logit_ale_lr = ALE(logit_fun_lr,
feature_names=feature_names,
target_names=target_names)
if dirname is not None:
logit_ale_lr.save(dirname)
return logit_ale_lr
def ale(model, frame: frame):
""" Calcula los valores de Accumulated Local Effect (ALE) del modelo para los datos en frame
Arg:
model: modelo entrenado con el cual calcular el partial dependence
frame: data de entrada del modelo
Returns:
Entrega un diccionario con los nombres de las variables como sus llaves y los valores una tupla de los valores que estas toman
y su valor ALE calculado
"""
exp_results = ALE(model.predict,
feature_names=frame.columns).explain(frame.values)
ale_results = {}
for i in range(len(exp_results['feature_names'])):
ale_results[exp_results['feature_names'][i]] = (
exp_results['feature_values'][i],
exp_results['ale_values'][i].reshape(
(exp_results['feature_values'][i].shape)))
return ale_results
def ale_explainer(iris_data, lr_classifier):
ale = ALE(predictor=lr_classifier.predict_proba,
feature_names=iris_data['metadata']['feature_names'])
return ale