bootstrap_CI.py

import numpy as np
from mlxtend.evaluate import bootstrap_point632_score
from sklearn.metrics import recall_score
from sklearn.datasets import make_classification
from sklearn.linear_model import LogisticRegression

# Create a dataset
X, y = make_classification(n_samples=1000, n_features=5, n_redundant=0)
# Add some noise to the features
X += np.random.normal(0, 2.5, X.shape)
# Use a Logistic regression classifier
clf = LogisticRegression(solver='lbfgs')
# Run the bootstrap simulation
scores = bootstrap_point632_score(clf, X, y, scoring_func=recall_score, method='.632')
# bootstrap estimate of the recall
print(np.mean(scores))
#percentile based 95% confidence interval
lower = np.percentile(scores, 2.5)
upper = np.percentile(scores, 97.5)
print(lower, upper)