Methods for quantifying the causal relationships in bivariate time series and review of performance

What is it about?

We have compared methods for estimating how much one time series 'influences' another, in a simple system containing two simultaneously observed components. We have investigated performance on a number of simulated datasets, as well as providing tests for their robustness to problems with data.

Featured Image

Photo by Chris Liverani on Unsplash

Photo by Chris Liverani on Unsplash

Why is it important?

Understanding cause-and-effects in observational time series data is a difficult challenge, involving contributions from different areas of mathematics. There is not an easy one-size-fits all solution to this problem, which requires careful consideration of other additional potential causes or confounding variables. Our work is a step towards this goal, via a greater understanding of which tools (and when) are suitable for quantifying relationships in a bivariate system.