sona-oversampling 0.1.3

Stochastic directional Oversampling using Negative Anomalous scores (SONA)

SONA

Stochastic directional Oversampling using Negative Anomalous scores for imbalanced dataset

Installation

To install SONA, type the following command in the terminal

pip install sona-oversampling            # normal install
pip install --upgrade sona-oversampling  # or update if needed

Required Dependencies :

• Python 3.9 or higher
• numpy>=2.0.2
• scipy>=1.8.1

SONA

The SONA function generates synthetic samples for the minority class by identifying border areas between classes and creating new points within a calculated safety radius.

SONA(X,y, min_label, new_label = 0)

Parameter	Type	Description
X	numpy.ndarray	The input feature matrix of shape $(n_{samples}, n_{features})$.
y	numpy.ndarray	The target labels.
min_label	int	The specific label identified as the minority class.
new_label	int	An offset added to min_label for the new synthetic samples (default is 0).

Returns

Parameter	Type	Description
X_augmented	numpy.ndarray	containing the original features plus the new synthetic samples.
y_augmented	numpy.ndarray	containing the original labels plus labels for the synthetic samples.

Usage Example

from sona_oversampling import SONA
from sklearn.datasets import make_circles, make_moons, make_blobs

# Generate 'Double circle' dataset with imbalance
X_circles, y_circles = make_circles(n_samples=(500, 100), noise=0.05, random_state=42)

# Generate 'Blue-moons' dataset with imbalance
X_moons, y_moons = make_moons(n_samples=(500, 100), noise=0.1, random_state=42)

# Generate imbalanced 'Gaussian cluster' dataset
X_blobs, y_blobs = make_blobs(n_samples=[500, 50], cluster_std=0.5, centers=[[0, 0], [1, 1]], random_state=42)

synthetic_datasets = [
    ("Double circle", (X_circles, y_circles)),
    ("Blue-moons", (X_moons, y_moons)),
    ("Gaussian cluster", (X_blobs, y_blobs))
]

for name, (X_original, y_original) in synthetic_datasets:
  
  X_oversampled, y_oversampled = SONA(X_original, y_original, min_label= 1, new_label=1)
  mask_maj = (y_oversampled == 0)
  mask_min_orig = (y_oversampled == 1)
  mask_min_syn = (y_oversampled == 2)

  plt.figure(figsize=(10, 8))

  # Plot Majority Class
  plt.scatter(X_oversampled[mask_maj, 0], X_oversampled[mask_maj, 1],
              c='grey', label='Majority Class (0)', alpha=0.5, s=20)

  # Plot Original Minority Class
  plt.scatter(X_oversampled[mask_min_orig, 0], X_oversampled[mask_min_orig, 1],
              c='blue', label='Original Minority (1)', s=30, edgecolors='k')

  # Plot Synthetic Minority Class
  plt.scatter(X_oversampled[mask_min_syn, 0], X_oversampled[mask_min_syn, 1],
              c='red', label='Synthetic Samples (2)', marker='x', s=40, alpha=0.8)

  plt.title(f"SONA Oversampling: {name} Dataset", fontsize=14)
  plt.xlabel("Feature 1")
  plt.ylabel("Feature 2")
  plt.legend(loc='best')
  plt.grid(True, linestyle='--', alpha=0.6)
  plt.axis('equal')

  plt.show()

Output