Suppose I have groups of event strings that I can easily self-correlate through having a specified pool size from which these event strings can be constructed from. For example,
I need 10 event strings and to have correlation = 2, I would construct a pool of 5 event strings to choose from. Then I would have 10 strings and 5 unique strings that were chosen from this pool. (I can add noise so that the correlation is more physical, etc.)
Let's say now I have many groups of these self-correlated event strings. I want to correlate these groups with each other.
So I need some sort of "correlation matrix" which is really only a triangular matrix, relating each group to another group.
Question: My issue is that I'm having trouble wrapping my head around how I would go about constructing "shared pools" between the *i*th and *j*th groups while making sure the other cross-correlations are maintained? Thanks!