A comprehensive machine learning toolkit that combines AutoML, genetic programming, clustering, and model explainability features for advanced data analysis and model development.

reveng/
├── src/                   # Core modules
│   ├── automl.py          # H2O AutoML implementation
│   ├── explainautoml.py   # Model explainability with LIME
│   ├── geneticlearn.py    # Genetic programming using DEAP
│   ├── h2owrapper.py      # H2O wrapper utilities
│   ├── preclustering.py   # DBSCAN clustering and outlier detection
│   └── preprocessing.py   # Data preprocessing pipeline
├── utils/                 # Utility functions
├── config/                # Configuration files
├── data/                  # Sample datasets
├── __main__.py            # Main entry point
└── __init__.py            # Package initialization

Configure your dataset in config/config.json (see Configuration section)

python -m reveng

Available commands:

• cl - Compute preclustering and outlier detection
• gp - Run genetic programming
• automl - Execute AutoML pipeline
• codb - Connect to database
• q - Quit the application

• AutoML Integration: Automated machine learning using H2O framework with support for multiple algorithms (Random Forest, GLM, Deep Learning, GBM)
• Genetic Programming: Feature selection and model generation using DEAP (Distributed Evolutionary Algorithms in Python)
• Clustering & Outlier Detection: DBSCAN-based clustering with automatic outlier identification
• Model Explainability: LIME (Local Interpretable Model-agnostic Explanations) integration for model interpretability
• Data Preprocessing: Comprehensive preprocessing pipeline with encoding, scaling, and feature selection
• Interactive CLI: Command-line interface for easy interaction with all features
• Configurable: JSON-based configuration system for flexible dataset handling

Edit config/config.json to configure your dataset:

{
  "your_dataset": {
    "path": "/data/your_file.csv",
    "target_output": "target_column_name",
    "exclude": ["target_column_name", "id_column"],
    "categorical_features": [0, 1, 4, 5],
    "inputing_features_gp": [3, 13, 14]
  },
  "automl_param": {
    "max_models": 10,
    "nfolds": 5,
    "seed": 42,
    "include_algos": ["DRF", "GLM", "DeepLearning", "GBM"]
  }
}

Configuration Parameters:

• path: Path to your CSV dataset
• target_output: Name of the target/output column
• exclude: List of columns to exclude from training (should include target column)
• categorical_features: List of column indices that contain categorical data
• inputing_features_gp: List of column indices to use for genetic programming
• automl_param: H2O AutoML configuration parameters

1. Configure your dataset in config/config.json
2. Run the application: python -m reveng
3. Execute AutoML: automl
4. View model performance and explanations

1. Run clustering: cl
2. View outlier detection results
3. Analyze cluster distributions

1. Run genetic programming: gp
2. View evolved feature combinations
3. Analyze fitness scores

• Python 3.7+
• Java 8+ (required for H2O)

Install the required packages:

pip install h2o pandas numpy scikit-learn deap lime matplotlib pyfiglet

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

For issues and questions, please open an issue on the GitHub repository.