spacebench

SpaCE, the Spatial Confounding Environment, loads benchmark datasets for causal inference methods tackling spatial confounding

The Spatial Confounding Environment loads benchmark datasets for causal inference that incorporartes spatial confounding. The SpaCE datasets contain real confounder and exposure/treatment data inspired by environmental health studies. The synthetic outcome and counterfactuals are generated for causal evaluation. They mimick real outcome data distributions learned with machine learning and neural network methods. Spatial confounding is achieved by masking influential confounders in the learned model.

Introduction

• About

Researchers

• Setting Up Environment
• Quickstart
  • Available datasets
  • Semi-synthetic data environment
• Modules
  • DataMaster Module
  • Environment Module
  • Evaluation Module

Community

• Contact Us
• spacebench (developing)
  • Added
  • Changed

Indices and tables

• Index

• Module Index

• Search Page