Skip to main content

About Xplainable

Overview​

xplainable is a Python package that leverages explainable machine learning for fully transparent machine learning and advanced data optimisation in production systems.

xplainable bridges the gap between data scientists, analysts, developers, and business domain experts by providing a simple Python API and web interface to manage machine learning systems at both a technical and managerial level. It achieves this by providing a set of tools that allow users to:

  • Quickly preprocess data and generate features
  • Train and evaluate machine learning models
  • Visualise and interpret model performance
  • Explain model predictions
  • Deploy models to production as REST APIs in seconds
  • Collaborate with other users on model development and evaluation
  • Save and load preprocessing pipelines and models across your team or organisation
  • Share model profiles with other users via xplainable cloud

At the core of xplainable is a set of novel explainable machine learning algorithms designed to provide similar performance to black box models while maintaining complete transparency.

These docs contain details about how xplainable works and how to get the most out of it.

Who is Xplainable For?​

xplainable is for anyone who wants to build machine learning models that can be easily understood and explained. Experienced professionals, novices, students, and hobbyists can all use the package given appropriate data. The only requirement is a basic understanding of Python and machine learning at a conceptual level.

The users who will get the most out of xplainable include:

  • Data scientists
  • Data analysts
  • Data engineers
  • Developers
  • Business domain experts

The package, combined with the web application, is designed to be used by individuals and teams within data-centric organisations.

Skill Requirements​

Anyone with a basic understanding of Python and machine learning at a conceptual level can use xplainable. The package is intuitive and easy to use by design, and the web application provides a simple interface for managing models and deployments.

Experienced Users​

Experienced users can use xplainable like any other open-source machine learning package. The package provides a simple API for training and evaluating models with the added benefit of novel model-tuning methods and advanced explainability tools.

These users can still benefit from the GUI tools of xplainable by streamlining the process of training and evaluating models, but they can also go as low-level into the code as they require for complete control over the model development process.

Novice Users​

Novice users can use xplainable to learn about machine learning and experiment with models and datasets using AutoML. The package provides a simple embedded GUI for training and evaluating models without having to write much code or understand the underlying algorithms.

xplainable also gives users with little to no experience with machine learning the ability to deploy models to production as REST APIs in seconds, which significantly reduces the barrier to entry of adding tangible value to data-centric organisations.

As novice users become more experienced, they can start to interface with xplainable at a lower level and start to use the more advanced features of the package.