HCrystal Ball

A library that unifies the API for most commonly used libraries and modelling techniques for time-series forecasting in the Python ecosystem.

The library consists of two main parts:

1. Wrappers - which bring different 3rd party libraries to time series compatible sklearn API

2. Model Selection - to enable gridsearch over wrappers, general or custom made transformers and add convenient layer over whole process (access to results, plots, storage, …)

Quick Installation:

Install HCrystalBall from pip

pip install hcrystalball

For full installation notes, see Installation

Note

For questions and contributions, please use

Contents:

• Installation
  • Install core HCrystalBall from pip or from conda-forge
  • Install other libraries you want to wrap
  • For parallel execution
  • Typical Installation
• Glossary
• Data Format
  • Model wrappers
  • Model selection
• Model Selection
  • Plotting
  • Parallel execution
  • Predefined parameter grid
  • How train-test split works
• Wrappers
  • Most wrappers
  • Sklearn compatible regressors

Quickstart

Tutorial

Logging best practices

API:

• API
  • hcrystalball.model_selection Package
  • hcrystalball.wrappers Package
  • hcrystalball.ensemble Package
  • hcrystalball.feature_extraction Package
  • hcrystalball.compose Package
  • hcrystalball.preprocessing Package
  • hcrystalball.metrics Package
  • hcrystalball.utils Module

Meta:

• Contribution
  • Development installation
  • Building documentation
  • Pre-commit hooks
  • Running tests
  • Creating new release
• Changelog
  • Version 0.1.9
  • Version 0.1.8
  • Version 0.1.7
  • Version 0.1.6
  • Version 0.1.5
  • Version 0.1.4
  • Version 0.1.3
  • Version 0.1.0

Indices and tables:

• Index

• Module Index

• Search Page