Home

Causal Inference on Time Series

We are investigating causal inference on time series. Therefore, we implemented the approach found in Chen's paper as well as a basic time series data generator in Python. In this implementation, the time lag that needs to be considered for the causal inference is estimated. We aim to find and compare the performance of different methods for estimating the relevant time lag and inferring the causal relationships within time series on artificial and real-world data. Based on Chen's approach, we have conceived an iterative algorithm to discover causal structures in time series fulfilling some constraints that works without explicitly estimating the time lag. This iterative approach is documented seperately.

Never forget to take notes and make assumptions explicit!