I realized for bDCM and DCM that Volterra Kernels are not derived for mixed observations (i.e. y matrices containing NaNs).
Using VBA_nanvar in VBA_getVolterraKernels could remove this issue (see pull request #71). But I do not know if that leads to an invalid derivation of kernels...
Dear all,
I realized for bDCM and DCM that Volterra Kernels are not derived for mixed observations (i.e. y matrices containing NaNs). Using VBA_nanvar in VBA_getVolterraKernels could remove this issue (see pull request #71). But I do not know if that leads to an invalid derivation of kernels...
Best, Simon