The MVGC Multivariate Granger Causality toolbox (Version 2) for Granger-causal inference from time-series data. CURRENTLY IN EARLY DEVELOPMENT - UNTESTED!
I want to use the MVGC toolbox to calculate Granger causality, but the program reports an error: ERROR: unstable VAR (unit root).
The reason is that the spectral radius calculated by the function var_specrad() is 1.658065, which is greater than 1. Does this mean that the data is non-stationary? I tried to use difference to make the data stationary, but the spectral radius is still greater than 1. What should I do to continue calculating Granger causality?(But I tried to use python code to do ADF and KPSS tests, and the results all showed that the data is stationary.)
I am a beginner in this area, can you help me solve this problem? Thank you very much!