Structure Uncertainty in Causal Inference
MCML Authors
David Strieder
Abstract
David Strieder
Abstract
In order to draw causal conclusions from available data, it is crucial to reason about the underlying causal structure that governs the data-generating process. In this publication-based thesis, we tackle the challenge of rigorously accounting for uncertainty in this underlying causal structure in causal inference. We present a framework based on test inversions to construct calibrated confidence regions for total causal effects that capture both sources of uncertainty: causal structure and numerical size of nonzero effects.
BibTeXKey: Str25