I am trying to use Granger causality for testing causal dependency between two variables. However most Granger causal models (and also the Vector Autoregression models) use single time series (one sample per variable) for each of the variables (or if there are multiple variables, then one time series for each variable). The Python statsmodels Granger causality implementation gives an example of this.
However in my case I have two variables, but multiple samples of time series for each variable and I want to fuse them into the Granger causal model together. That is instead of considering each time series pair separately for the variables for the output of the Granger causal model, is it possible to use all of them together to form some kind of model analogous to the regression model.
