Using Unsupervised Learning to Help Discover the Causal Graph

TL;DR

AitiaExplorer is an unsupervised learning-based tool designed to facilitate causal discovery by automatically selecting important features and generating candidate causal graphs for review.

Contribution

The paper introduces AitiaExplorer, a novel software that uses unsupervised learning for feature selection to improve causal graph discovery.

Findings

AitiaExplorer effectively selects relevant features from datasets.

The tool generates plausible causal graph candidates.

It meets the outlined requirements for exploratory causal analysis.

Abstract

The software outlined in this paper, AitiaExplorer, is an exploratory causal analysis tool which uses unsupervised learning for feature selection in order to expedite causal discovery. In this paper the problem space of causality is briefly described and an overview of related research is provided. A problem statement and requirements for the software are outlined. The key requirements in the implementation, the key design decisions and the actual implementation of AitiaExplorer are discussed. Finally, this implementation is evaluated in terms of the problem statement and requirements outlined earlier. It is found that AitiaExplorer meets these requirements and is a useful exploratory causal analysis tool that automatically selects subsets of important features from a dataset and creates causal graph candidates for review based on these features. The software is available at https://github.com/corvideon/aitiaexplorer