Home

Welcome to the Ensemble Genetic Algorithm project wiki!

This project provides a robust framework for evolving machine learning ensembles for binary classification problems using genetic algorithms. It’s designed to be extensible, highly configurable, and flexible, allowing users to customize base learners, genetic algorithm hyperparameters, and feature space exploration.

Key Features

• Genetic Algorithm Optimization: Utilizes a genetic algorithm to search for optimal ensembles of machine learning classifiers.

• Extensible Base Learners: Easily integrate various machine learning algorithms as base learners.

• Configurable Hyperparameters: Fine-tune genetic algorithm parameters (e.g., population size, generations) and base learner configurations.

• Feature Space Exploration: Supports grid search over feature spaces and feature transformations.

• GPU Acceleration: Optional GPU support for PyTorch-based models.

• Comprehensive Setup: Includes a setup.sh script for automated environment creation and dependency installation.

Getting Started

To get started with the project, please refer to the Installation Guide.

New in v1.0+: Check out our new API Reference for comprehensive documentation of all public classes and methods.

Documentation

Explore the following wiki pages for detailed information:

• Getting Started

  • Installation Guide: How to set up your development environment (Python >=3.12).

  • Usage Guide: How to run experiments using main.py and config.yml.

  • Data Preparation Guide: The required format for your input data.

• Core Concepts

  • Architectural Overview: A high-level look at the project’s components.

  • Technical Deep Dive: In-depth technical details and implementation architecture (includes performance benchmarks).

  • Configuration Guide: A detailed guide to the config.yml file.

  • Hyperparameter Reference Guide: A reference for all configurable parameters.

• Guides & Tutorials

  • Interpreting Experiment Results: How to understand the plots and outputs.

  • Evaluating Final Models with EnsembleEvaluator: How to validate your final models on unseen data.

  • Model Deployment Guide: How to export and serve your final models.

  • Adding a New Base Learner: How to extend the project with new models.

  • Best Practices and Tips: Tips for running experiments effectively.

• Reference

  • API Reference: Complete API reference with all public classes and methods.

  • Troubleshooting Guide: Solutions for common errors.

  • Project Structure: An overview of the repository’s file layout.

License

This project is licensed under the MIT License.

---