Asymptotics of \(k\)-mean clustering under non-i.i.d. sampling (Q1897095)

The asymptotics of \(k\)-mean clustering under two sets of relaxed sampling assumptions are reported. The first set is that the observations are independent but come from different distributions. For this set of assumptions, substantial modifications of the regularity conditions lead only to a consistency result. The second set of sampling assumptions is that the observations form a stationary mixing process, either strongly- mixing or \(\varphi\)-mixing. Under these assumptions only minor modifications of the regularity conditions are needed in order that the entire asymptotic theory carries over. This case includes autoregressive processes.