11
0

Asymmetric predictability in causal discovery: an information theoretic approach

Abstract

Causal investigations in observational studies pose a great challenge in research where randomized trials or intervention-based studies are not feasible. We develop an information geometric causal discovery and inference framework of "predictive asymmetry". For (X,Y)(X, Y), predictive asymmetry enables assessment of whether XX is more likely to cause YY or vice-versa. The asymmetry between cause and effect becomes particularly simple if XX and YY are deterministically related. We propose a new metric called the Directed Mutual Information (DMIDMI) and establish its key statistical properties. DMIDMI is not only able to detect complex non-linear association patterns in bivariate data, but also is able to detect and infer causal relations. Our proposed methodology relies on scalable non-parametric density estimation using Fourier transform. The resulting estimation method is manyfold faster than the classical bandwidth-based density estimation. We investigate key asymptotic properties of the DMIDMI methodology and a data-splitting technique is utilized to facilitate causal inference using the DMIDMI. Through simulation studies and an application, we illustrate the performance of DMIDMI.

View on arXiv
Comments on this paper