In StatsBase.crosscor!()
why is the standard deviation for normalisation calculated before the lag is applied to the input vectors, rather than after?
This doesn’t appear to follow a standard correlation formula.
Cross-Correlation with StatsBase crosscor answer does not make sense