imblearn_random_oversampler_balance_imbalanced_dataset.py

1from sklearn.datasets import make_classification
2from imblearn.over_sampling import RandomOverSampler
3from collections import Counter
4
5# Generate a synthetic imbalanced dataset
6X, y = make_classification(n_samples=5000, n_features=2, n_informative=2,
7                           n_redundant=0, n_repeated=0, n_classes=3,
8                           n_clusters_per_class=1,
9                           weights=[0.01, 0.05, 0.94],
10                           class_sep=0.8, random_state=0)
11
12print(f'Original dataset shape {Counter(y)}')
13
14# Instantiate the sampler
15ros = RandomOverSampler(random_state=0)
16
17# Resample the dataset
18X_resampled, y_resampled = ros.fit_resample(X, y)
19
20print(f'Resampled dataset shape {Counter(y_resampled)}')