I am unsure if an algorithm exists to do this:
I have an alphabet {a,b,c,d,e,f} which I have many permutations of. e.g.
abfdce cedbaf dacfeb
and i know that on average 'a' occurs in the first position with probability x, 'b' occurs with probability y etc. so i have aggregate probabilities about which letter occurs in which position. How do I draw new permutations such that, on average, they have the same statistics as the data i have? My alphabet size is ~50 so i can't enumerate them all or anything like that.
[link] [2 comments]