vfa 0.1.0

Variance Feature Analysis for binary classification feature selection

VFA - Variance Feature Analysis

A Python package for binary classification feature selection using variance-based analysis.

Overview

Variance Feature Analysis (VFA) implements a feature selection method based on the Class-Variance Ratio (CVR). It selects the most discriminative features for binary classification tasks by analyzing the ratio of between-class variance to total variance.

Installation

pip install vfa

Features

• Fast variance-based feature selection for binary classification
• Automatic feature ranking using Class-Variance Ratio (CVR)
• Weighted feature aggregation
• Compatible with scikit-learn workflows
• Lightweight with minimal dependencies (only NumPy)

Usage

from vfa import variance_feature_analysis
import numpy as np

# Example data
X = np.random.rand(100, 20)  # 100 samples, 20 features
y = np.random.randint(0, 2, 100)  # Binary labels

# Select top 5 features
X_selected, f_aggregated, selected_indices, scores = variance_feature_analysis(X, y, k=5)

print(f"Selected feature indices: {selected_indices}")
print(f"Feature scores: {scores[selected_indices]}")

Parameters

• X (array-like): Training data features of shape (n_samples, n_features)
• y (array-like): Target labels of shape (n_samples,) - must be binary
• k (int, default=8): Number of top features to select
• epsilon (float, default=1e-12): Small constant to prevent division by zero

Returns

• X_selected: Selected feature subset
• f_aggregated: Weighted aggregation of selected features
• selected_idx: Indices of selected features
• scores: CVR scores for all features

How It Works

The algorithm:

1. Computes within-class and between-class variance for each feature
2. Calculates the Class-Variance Ratio (CVR) = B / (B + W)
3. Selects the top-k features with highest CVR scores
4. Returns selected features and their weighted aggregation

Requirements

• Python >=3.8
• NumPy >=1.20.0

Development

# Clone the repository
git clone https://github.com/nqmn/vfa.git
cd vfa

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest

License

MIT License - see LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Citation

If you use this package in your research, please cite:

@software{vfa2024,
  title={VFA: Variance Feature Analysis},
  author={Mohd Adil Mokti},
  year={2026},
  url={https://github.com/nqmn/vfa}
}