Integer-valued Autoregressive Moving Average Models
ABSTRACT:
A stochastic process X(t) is INAR(1) if there exists a sequence Z(t)
of iid non-negative integer-valued random variables and a sequence
Y(t,i) of iid Bernoulli random variables independent of Z(t) such that
X(t)=sum{i=1..X(t-1)}Y(t,i)+Z(t)
This process has many properties similar to the AR(1) process. It has
extensions to INARMA(p,q) processes. The INAR(1) process also has branching
process and queueing interpretations. Some properties of such
processes will be discussed.
About the Speaker
Information about the speaker will be posted here shortly
References
Al-Osh, M. A. and Alzaid, A. (1987) First order integer-valued autoregressive (INAR(1))
J. Time Series Anal. 8, 261-275.
Al-Osh, M. A. and Alzaid, A. (1988). Integer-valued moving-average processes. Statist. Papers, 29, 281-300.
Alzaid, A. and Al-Osh, M. A. (1990) An integer-valued pth order autoregressive structure (INAR(P)) process. J. Appl. Prob. 27, 314-324.
Du, J. G. and Li, Y (1991) The integer-valued autoregressive (INAR(P)) model. J. Time
Series Anal. 12, 129-142.
Latour, A (1997) The multivariate GINAR(p) process. Adv. Appl. Prob. 29,228-247
McKenzie, E. (1986) Autoregressive moving-average processes with negative-binomial and geometric marginal distributions. Adv. Appl. Prob. 18, 679-705
McKenzie, E. (1988) Some ARMA models for dependent sequences of Poisson counts. Adv. Appl. Prob. 20, 822-835.
Park, Y. and Kim, M. (1997) Some basic and asymptotic properties in INMA(q) processes. J. Korean Statist. Soc. 26, 155-170.