![](https://csdnimg.cn/release/download_crawler_static/12141122/bg2.jpg)
varying-coefficient partially linear regression models. More references and techniques can be
found in the monograph Owen (2001).
Integer-valued time series data occur in many fields including insurance, medicine, queueing
systems, communications, reliability theory and meteorology. Examples of such data include
the data on the number of patients in a hospital at a specific time, the number of bases of DNA
sequences, the number of road accidents in a certain area in successive months, and the number
of machines waiting for maintenance. In recent years, a number of authors have studied the
problem of integer-valued time series analysis and most models have been proposed based on
the thinning operator “ ◦ ” due to Steutel and van Harn (1979). Let X be a non-negative
integer-valued random variable and α ∈ [0, 1). Then “ ◦ ” is defined as
α ◦X =
X
X
i=1
B
i
,
where {B
i
} is an independent and identically distributed (iid) Bernoulli random sequence with
P (B
i
= 1) = 1 −P (B
i
= 0) = α that is independent of X. Some processes have been discussed
by McKenzie (1986, 1988), Al-Osh and Alzaid (1987), Alzaid and Al-Osh (1988) and Al-Osh
and Aly (1992), and Zheng et al. (2006, 2007) introduced random coefficient integer-valued
autoregressive models.
In this paper, an EL method is developed for the pth-order random coefficient integer-
valued autoregressive (RCINAR(p)) process. In Section 2, we review the definition of the
RCINAR(p) process and propose the empirical likelihood ratio (ELR) statistic. In Section 3,
the asymptotic distribution of the ELR statistic is established and the confidence regions for
the parameter of interest are derived. Section 4 presents some simulation results.
1 Empirical likelihood for the RCINAR(p) process
The RCINAR(p) process {X
t
} satisfies an equation of the form:
X
t
=
p
X
i=1
φ
(t)
i
◦ X
t−i
+ Z
t
, t ≥ 1, (1)
where {φ
(t)
i
} is an iid sequence with a cumulative distribution function P
φ
i
on [0,1) and {Z
t
}
is an iid non-negative integer-valued sequence with a probability mass function f
z
> 0. Also
in the above, {φ
(t)
i
, 1 ≤ i ≤ p} and {Z
t
} are independent sequences. Let λ = E(Z
t
), φ
i
=
E(φ
(t)
i
), i = 1, 2, ··· , p, and we suppose
p
P
i=1
φ
i
< 1.
Now we introduce the EL method. Let x
1
, ··· , x
n
be iid discrete-valued d-dimensional
random vectors with common cumulative distribution function F and mean µ. The EL function
is
L(F ) =
n
Y
t=1
dF (x
t
) =
n
Y
t=1
ω
t
(2)
- 2 -
中国科技论文在线
http://www.paper.edu.cn
评论0
最新资源