svdynamics 0.0.0a0

Support Vector Dynamics: mixed (composite) kernels for scikit-learn SVMs.

Welcome to svdynamics! Support Vector Dynamics is a lightweight, scikit-learn compatible Python library for building and using mixed (composite) kernels for support vector machines. It provides a simple and extensible interface for combining multiple kernel functions into a single weighted kernel, while remaining fully compatible with existing sklearn pipelines, cross-validation, and calibration workflows.

svdynamics focuses on making kernel composition a first-class modeling primitive for both classification and regression, without requiring any changes to the underlying scikit-learn API.

Prerequisites

Before you install svdynamics, ensure your system meets the following requirements:

• Python: Version 3.8 or higher.

Additionally, svdynamics depends on the following packages, which will be automatically installed:

• numpy: version 1.21 or higher
• scikit-learn: version 1.0 or higher

💾 Installation

To install svdynamics, simply run the following command in your terminal:

pip install svdynamics

Quick Start

import numpy as np
from sklearn.datasets import make_classification
from svdynamics import CompositeKernel, SVDClassifier

X, y = make_classification(n_samples=300, n_features=10, random_state=0)

kernel = CompositeKernel(
    kernels=[
        ("rbf", {"gamma": 0.2}),
        ("linear", {}),
        ("poly", {"degree": 2, "coef0": 1.0}),
    ],
    weights=[0.6, 0.3, 0.1],
    normalize=True,
)

clf = SVDClassifier(C=1.0, kernel=kernel, probability=True, random_state=0)
clf.fit(X, y)
proba = clf.predict_proba(X[:5])
pred = clf.predict(X[:5])

print(proba)
print(pred)

📄 Official Documentation

https://lshpaner.github.io/svdynamics

🌐 Authors' Website

1. Leon Shpaner

⚖️ License

svdynamics is distributed under the MIT License. See LICENSE for more information.