pylwl 0.1.0

PyLWL: A Python Framework for Locally Weighted Learning

PyLWL: A Python Framework for Locally Weighted Learning

---

📌 Overview

PyLWL is an open-source Python library that provides a unified, extensible, and user-friendly implementation of Locally Weighted Learning (LWL) algorithms for supervised learning. It implements differentiable and gradient-descent-based local models for both classification and regression tasks.

Features

• 📌 GdLwClassifier: Local weighted classifier using logistic regression with support for binary and multiclass classification.
• 📌 GdLwRegressor: Local weighted regressor using linear regression optimized with MSE loss.
• 📌 LwClassifier and LwRegressor: Local weighted classifier/regressor with a fixed kernel.
• 🧠 Supports any differentiable kernel function (e.g., Gaussian, Epanechnikov).
• ⚙️ Built with PyTorch, and fully compatible with Scikit-Learn pipeline and metrics.
• 🔧 Configurable optimizer (Adam, SGD, etc.) and hyperparameters.
• 🔍 Built-in support for model evaluation and scoring.

📦 Installation

You can install the library using pip (once published to PyPI):

pip install pylwl

After installation, you can import PyLWL as any other Python module:

$ python
>>> import pylwl
>>> pylwl.__version__

🚀 Quick Start

Classification

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from pylwl import LwClassifier

# Load data
X, y = load_iris(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)

# Train LVQ1 model
model = LwClassifier(kernel="gaussian", tau=1.0)
model.fit(X_train, y_train)

# Evaluate
y_pred = model.predict(X_test)
print("Accuracy:", model.score(X_test, y_test))

Regression

from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import train_test_split
from pylwl import LwRegressor

X, y = fetch_california_housing(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)

reg = LwRegressor(kernel='gaussian', tau=1.0)
reg.fit(X_train, y_train)
print("R2 score:", reg.score(X_test, y_test))

Please read the examples folder for more use cases.

📚 Documentation

Documentation is available at: 👉 https://pylwl.readthedocs.io

You can build the documentation locally:

cd docs
make html

🧪 Testing

You can run unit tests using:

pytest tests/

🤝 Contributing

We welcome contributions to PyLWL! If you have suggestions, improvements, or bug fixes, feel free to fork the repository, create a pull request, or open an issue.

📄 License

This project is licensed under the GPLv3 License. See the LICENSE file for more details.

Citation Request

Please include these citations if you plan to use this library:

@software{thieu20250517PyLWL,
  author       = {Nguyen Van Thieu},
  title        = {PyLWL: A Python Framework for Locally Weighted Learning},
  month        = may,
  year         = 2025,
  doi         = {10.6084/m9.figshare.29089784},
  url          = {https://github.com/thieu1995/PyLWL}
}

Official Links

• Official source code repo: https://github.com/thieu1995/PyLWL
• Official document: https://pylwl.readthedocs.io/
• Download releases: https://pypi.org/project/pylwl/
• Issue tracker: https://github.com/thieu1995/PyLWL/issues
• Notable changes log: https://github.com/thieu1995/PyLWL/blob/master/ChangeLog.md
• Official chat group: https://t.me/+fRVCJGuGJg1mNDg1

---

Developed by: Thieu @ 2025