Hey all, I read in a paper about the zipf distribution, which seems different from what i learned here. It says: we can generate a random dataset of pairs of the form (x,y). By modifying some variable Z during data generation, we can generate dataset where x and y can be closely correlated or totally independent. I am wondering whether it possible to achieve that. I want to generate random pairs of (x,y), where I can control the correlation between x and y. Thanks a lot, Senlin [[alternative HTML version deleted]]