I am working on a project in which I use the user-written R package bvarsv (here an example were this package is being used) to implement a time-varying-parameter analysis. I have not found a way to test if my model is stable, or to test my residuals for normality or auto-correlation. Is there a standard way of running these tests - if not for the model in this specific package, but for Bayesian VARs in general?
Maybe I am also missing something here and these tests are actually not necessary (to my knowledge, for example, the 2005 Primiceri paper does not do any such tests), if this is the case than I would be happy for an explanation why.
与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…
