Causal Inference by Identification of Vector Autoregressive Processes with Hidden Components

By Philipp Geiger, Kun Zhang, Bernhard Schölkopf, Mingming Gong, Dominik Janzing. 2014

85 citations recorded by the source corpus.

A widely applied approach to causal inference from a time series X, often referred to as "(linear) Granger causal analysis", is to simply regress present on past and interpret the regression matrix B causally. However, if there is an unmeasured time series Z that influences X, then this approach can lead to wrong causal conclusions, i.e., distinct from those one would draw if one had additional information such as Z. In this paper we take a different approach: We assume that X together with some hidden Z forms a first order vector autoregressive (VAR) process with transition matrix A, and argue why it is more valid to interpret A causally instead of B. Then we examine under which conditions the most important parts of A are identifiable or almost identifiable from only X. Essentially, sufficient conditions are (1) non-Gaussian, independent noise or (2) no influence from X to Z. We present two estimation algorithms that are tailored towards conditions (1) and (2), respectively, and evaluate them on synthetic and real-world data. We discuss how to check the model using X.

Subfields: Artificial Intelligence, Control and Systems Engineering, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

In Statistics and Probability

Citations in the corpus

Causal Discovery and Forecasting in Nonstationary Environments with State-Space Models (2019, methodology): “As future work, we will extend our methods to cover nonlinear causal relationships, to partially observable processes, as studied in (Geiger et al., 2015), and to causal models with instantaneous cycles.”
A Survey of Learning Causality with Data (2018, background): “e, for learning causal relationship in time series data, hidden time series acts like unobserved confounders in i.i.d. data. Confounding bias can lead to wrong causal conclusions [131]. Geiger et al. [44] showed that causal relationships can be discovered under confounding bias with several assumpti”
A Survey of Learning Causality with Data: Problems and Methods (2018, background): “[44] showed that causal relationships can be discovered under confounding bias with several assumptions.”