Skip to content

Abstract Search

Primary Submission Category: Causal Discovery

Causal Discovery for Observational Categorical Data

Authors: Yang Ni,

Presenting Author: Yang Ni*

Causal discovery for quantitative data has been extensively studied but less is known for categorical data. I will present novel causal models for categorical data. For ordinal categorical data, our model is based on ordinal regression whereas, for nominal categorical data, it is based on a new classifier, termed classification with optimal label permutation. Under either causal model, we establish its causal identifiability property with observation data alone. Through experiments with synthetic and real data, we demonstrate the favorable performance of the proposed causal models compared to state-of-the-art methods.