Skip to content

Installation Guide

This guide will help you install skforecast, a powerful library for time series forecasting in Python. The default installation of skforecast includes only the essential dependencies required for basic functionality. Additional optional dependencies can be installed for extended features.

Python PyPI Conda

Basic installation

Skforecast requires Python 3.9 or higher. It is available on PyPI and can be installed using pip. You can also install it via conda from the conda-forge channel.

To install the basic version of skforecast with its core dependencies, run:

pip install skforecast

Specific version:

pip install skforecast==0.17.0

Latest (unstable):

pip install git+https://github.com/skforecast/skforecast@master

The following dependencies are installed with the default installation:

  • numpy>=1.24
  • pandas>=1.5
  • tqdm>=4.57
  • scikit-learn>=1.2
  • optuna>=2.10
  • joblib>=1.1
  • numba>=0.59
  • rich>=13.9

Optional dependencies

To install the full version with all optional dependencies:

pip install skforecast[full]

For specific use cases, you can install these dependencies as needed:

Sarimax

pip install skforecast[sarimax]
  • statsmodels>=0.12, <0.15

Plotting

pip install skforecast[plotting]
  • matplotlib>=3.3, <3.11
  • seaborn>=0.11, <0.14
  • statsmodels>=0.12, <0.15

Deep Learning

pip install skforecast[deeplearning]
  • keras>=3.0, <4.0
  • matplotlib>=3.3, <3.11