mlarena 0.1.9

An algorithm-agnostic machine learning toolkit for model training, diagnostics and optimization

MLArena

An algorithm-agnostic machine learning toolkit for model training, diagnostics and optimization.

Publications

Read about the concepts and methodologies behind MLArena through these articles:

1. Algorithm-Agnostic Model Building with MLflow - Published in Towards Data Science

   A foundational guide demonstrating how to build algorithm-agnostic ML pipelines using mlflow.pyfunc. The article explores creating generic model wrappers, encapsulating preprocessing logic, and leveraging MLflow's unified model representation for seamless algorithm transitions.

2. Explainable Generic ML Pipeline with MLflow - Published in Towards Data Science

   An advanced implementation guide that extends the generic ML pipeline with more sophisticated preprocessing and SHAP-based model explanations. The article demonstrates how to build a production-ready pipeline that supports both classification and regression tasks, handles feature preprocessing, and provides interpretable model insights while maintaining algorithm agnosticism.

Installation

The package is undergoing rapid development at the moment (pls see CHANGELOG for details), it is therefore highly recommended to install with specific versions. For example

pip install mlarena==0.1.9

If you are using the package in Databricks ML Cluster with DBR runtime >= 15.2, you can try installing without dependencies (experimental feature):

pip install mlarena --no-deps

Usage Example

• For quick start with a basic example, see 1.basic_usage.ipynb.
• For more advanced examples, see 2.advanced_usage.ipynb.
• For visualization utilities, see 3.utils_plot.ipynb.
• For handling common challenges in machine learning, see 4.ml_discussions.ipynb.

Visual Examples:

Model Performance Analysis

Explainable ML

One liner to create global and local explaination based on shap that will work across various classification and regression algorithms.

Hyperparameter Optimization

Parallel Coordinate plot for hyperparameter search space diagnostics.

Features

• Algorithm Agnostic ML Pipeline:

  • End-to-end workflow from preprocessing to deployment
  • Model-agnostic design (works with any scikit-learn compatible model), easily experiment with and swap between algorithms
  • Support for both classification and regression tasks
  • Early stopping and validation set support
  • MLflow integration for experiment tracking and deployment

• Intelligent Preprocessing:

  • Automated feature type detection and handling
  • Smart encoding recommendations based on feature cardinality and rare category
  • Target encoding with visualization to support smoothing parameter selection
  • Tunable drop options to optimize one-hot encoding based on model (tree vs linear) and feature type (binary vs multi-category)
  • Missing value handling with configurable strategies
  • Feature selection recommendations with mutual information analysis

• Advanced Model Evaluation:

  • Comprehensive metrics for both classification and regression
  • Diagnostic visualization of model performance
  • Threshold analysis for classification tasks
  • SHAP-based model explanations (global and local)
  • Cross-validation with variance penalty

• Hyperparameter Optimization:

  • Bayesian optimization with Hyperopt
  • Cross-validation based tuning
  • Parallel coordinates visualization for search space analysis
  • Early stopping to prevent overfitting
  • Variance penalty to ensure stable solutions

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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