Displaying 1 result from an estimated 1 matches for "dememorisation".
2007 Oct 02
1
permutations of a binary matrix with fixed margins
...a chain where each permutation is seeded with
the previous one.
Potential problems:
1. Does it really explore all possible permutations? Does it do it in an
unbiased way?
2. Related to above: there is potential autocorrelation (although I haven't
found it with my data set), so might need some dememorisation steps.
3. It's slow and dememorisation makes it slower.
4. It isn't clear to me whether it needs the added stochastic element, i.e.
squares are only flipped if "runif(1)<0.5". In practice it seems to work
without it (as far as I can tell, i.e. it isn't autocorrelated using...