ASYMPTOTICS OF CERTAIN CONDITIONALLY IDENTICALLY DISTRIBUTED SEQUENCES
PATRIZIA BERTI, EMANUELA DREASSI, LUCA PRATELLI, AND PIETRO RIGO
Abstract. Let S be a Borel subset of a Polish space and (X
n: n ≥ 1) a sequence of S-valued random variables. Fix a Borel probability measure σ
0on S, a constant q
0∈ [0, 1] and a measurable function q
i: S
i→ [0, 1] for each i ≥ 1. Suppose X
1∼ σ
0and
P X
n+1∈ · | X
1, . . . , X
n= σ
0n
Y
i=1
Q
i+ δ
Xn(1 − Q
n) +
n−1
X
i=1
δ
Xi(1 − Q
i)
n
Y
j=i+1
Q
jwhere Q
i= q
i−1(X
1, . . . , X
i−1). Sequences of this type, introduced in [10], are conditionally identically distributed and play a role in Bayesian predictive inference. This paper deals with the asymptotics of (X
n). As expected, (X
n) exhibits different behaviors depending on the Q
i. For instance, (X
n) converges a.s. if α ≤ Q
i≤ β a.s. for all i, where 0 < α ≤ β < 1 are constants, while (X
n) does not converge even in probability if σ
0is nondegenerate, Q
i> 0 for all i and P
i