# Causal Network Inference by Optimal Causation Entropy

## TL;DR

This paper introduces a mathematically grounded, information-theoretic approach called causation entropy for inferring causal networks from time series data, with proven optimality and improved efficiency over existing methods.

## Contribution

It develops the theory of causation entropy, establishes the optimal causation entropy principle, and creates efficient algorithms for causal network inference from time series data.

## Key findings

- The method outperforms conditioned Granger causality and transfer entropy.
- Inference accuracy depends on network density and information diffusion, not just number of nodes.
- The approach is validated on Gaussian processes and large random networks.

## Abstract

The broad abundance of time series data, which is in sharp contrast to limited knowledge of the underlying network dynamic processes that produce such observations, calls for a rigorous and efficient method of causal network inference. Here we develop mathematical theory of causation entropy, an information-theoretic statistic designed for model-free causality inference. For stationary Markov processes, we prove that for a given node in the network, its causal parents forms the minimal set of nodes that maximizes causation entropy, a result we refer to as the optimal causation entropy principle. Furthermore, this principle guides us to develop computational and data efficient algorithms for causal network inference based on a two-step discovery and removal algorithm for time series data for a network-couple dynamical system. Validation in terms of analytical and numerical results for Gaussian processes on large random networks highlight that inference by our algorithm outperforms previous leading methods including conditioned Granger causality and transfer entropy. Interestingly, our numerical results suggest that the number of samples required for accurate inference depends strongly on network characteristics such as the density of links and information diffusion rate and not necessarily on the number of nodes.