Nonlinear positive AR(2) processes. (English) Zbl 0714.62087

Summary: Let \(e_ t\) be a strict white noise such that \(e_ t>0\). Consider a process \(X_ t\) given by \[ X_ t=b_ 1h_ 1(X_{t-1})+b_ 2h_ 2(X_{t-2})+e_ t, \] where \(b_ 1\geq 0\), \(b_ 2\geq 0\) and \(h_ 1,h_ 2\) are nondecreasing positive functions. Let a realization \(X_ 1,...,X_ n\) be given. Denote \(b^*_ 1\), \(b^*_ 2\) the solution of the linear program \(\max (\beta_ 1+\beta_ 2)\) under the conditions \(\beta_ 1\geq 0\), \(\beta_ 2\geq 0\), \(\beta_ 1h_ 1(X_{t- 1})+\beta_ 2h_ 2(X_{t-2})\leq X_ t\) \((t=3,...,n)\). It is proved that \(b^*_ 1\) and \(b^*_ 2\) are strongly consistent estimators for \(b_ 1\) and \(b_ 2\), respectively.


62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH)
60G12 General second-order stochastic processes
90C05 Linear programming
90C90 Applications of mathematical programming
Full Text: DOI


[1] DOI: 10.1080/03610928808829693 · Zbl 0639.62082 · doi:10.1080/03610928808829693
[2] DOI: 10.1111/j.1467-9892.1989.tb00011.x · Zbl 0671.62090 · doi:10.1111/j.1467-9892.1989.tb00011.x
[3] DOI: 10.1080/03610928908830139 · Zbl 0696.62347 · doi:10.1080/03610928908830139
[4] DOI: 10.1080/02331888908802217 · Zbl 0685.62066 · doi:10.1080/02331888908802217
[5] Andisl J., Comment. Math. Univ. 30 pp 403– (1989)
[6] Andiil J., Kybemetika 24 pp 372– (1988)
[7] Bell C. B, Commun. Statist.-Theor. Meih 15 pp 2267– (198) · Zbl 0604.62087 · doi:10.1080/03610928608829248
[8] DOI: 10.1080/02331888708802055 · Zbl 0636.62087 · doi:10.1080/02331888708802055
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.