On the pitfalls of Gaussian likelihood scoring for causal discovery
Christoph Schultheiss, Peter Bühlmann · Journal of Causal Inference · 2023
Abstract We consider likelihood score-based methods for causal discovery in structural causal models. In particular, we focus on Gaussian scoring and analyze the effect of model misspecification in terms of non-Gaussian error distribution. We present a surprising negative result for Gaussian likelihood scoring in combination with nonparametric regression methods.