sega_learn 0.1.12

Sega Learn is a Python package for machine learning and data science.

SEGA_LEARN

SEGA_LEARN is a Python package for machine learning and data science, offering from-scratch implementations of various algorithms. It's designed primarily for educational purposes, allowing users to explore the mechanics of ML models, inspired by libraries like scikit-learn and PyTorch.

Core Features

• Wide Range of Algorithms: Implements models for:
  • Automated Machine Learning (AutoML)
  • Clustering (e.g., KMeans, DBSCAN)
  • Linear Models (e.g., OLS, Ridge, Lasso, Logistic Regression, LDA, QDA)
  • Nearest Neighbors (KNeighborsClassifier, KNeighborsRegressor)
  • Neural Networks (flexible architecture with NumPy, Numba, and CuPy backends)
  • Support Vector Machines (Linear, Generalized with Kernels, One-Class)
  • Tree-based Models (Decision Trees, Random Forests, Gradient Boosting, AdaBoost, Isolation Forest)
  • Time Series Analysis (ARIMA, SARIMA, Decomposition, Exponential Smoothing)
• Pipelines: Generic and forecasting pipelines for streamlining ML workflows.
• Utilities: A rich set of tools for:
  • Data preparation and preprocessing (scaling, encoding, imputation)
  • Model evaluation metrics
  • Model selection (Grid Search, Random Search)
  • Data augmentation (SMOTE, over/under-sampling)
  • Visualization and animation helpers

Installation

You can install SEGA_LEARN directly from PyPI:

pip install sega_learn

Quick Links

• Homepage & Repository: https://github.com/SantiagoEnriqueGA/sega_learn
• Full Documentation: https://santiagoenriquega.github.io/sega_learn/sega_learn
• Contributing/Development: https://github.com/SantiagoEnriqueGA/sega_learn/blob/main/DEVELOPMENT.md

License

This project is licensed under the MIT License - see the LICENSE.txt file on GitHub for details.