On Pareto process
Let \((V_n )_{n \geqslant 1} \) and \((\varepsilon _n )_{n \geqslant 1} \) be two independent sequences of i.i.d. random variables such that \(P(V_1 < v) = v^\alpha ,\;\alpha > 0,\quad v < 1.\) Define a sequence \((X_n )_{n \geqslant 1} \) by \(X_0 = \varepsilon _1 \) and \(X_n = V_n^{ - 1} \min (X_{n - 1} ,\varepsilon _n ),\;n \geqslant 1.\) It is proved that \((X_n )_{n \geqslant 1} \) is stationary if and only if \(\varepsilon _1 \) has a Pareto distribution with parameters \(\sigma > 0\) and \(\alpha \), that is, \( P(\varepsilon _1 > x) = (1 + (x / \sigma )^\alpha )^{ - 1},\;\;x \geqslant 0. \) This result allows the author to define and study the corresponding AR(1) Pareto process as well as a first order autoregressive moving average (ARMA (1,1)) Pareto process and a multivariate AR(1) Pareto process. For the entire collection see [Zbl 1051.68019].
- A characterization of the Pareto process among stationary stochastic processes of the form \(X_ n=c\,\min (X_{n-1},Y_ n)\)
- A class of stationary Markov processes
- Pareto processes
- Pareto processes
- scientific article; zbMATH DE number 4109901 (Why is no real title available?)
- Autoregressive logistic processes
- Extremal and additive processes generated by Pareto distributed random vectors
- On Pareto's Law
This page was built for publication: On Pareto process
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3157527)
