scientific article; zbMATH DE number 16828

From MaRDI portal
Publication:3973518





[For the entire collection see Zbl 0722.00031.] The author examines the properties of processes whose joint distributions are Eyraud-Farlie-Gumbel-Morgenstern distributions, i.e. the joint distribution function of \(X_{t_ 1},X_{t_ 2},\dots,X_{t_ n}\) has the form \[ F(x_ 1,x_ 2,\dots,x_ n)=\left\{\prod_{i=1}^ n F_ i(x_ i)\right\}\left\{1+\sum_{1\leq k}\sum_{<j\leq n}\alpha(t_ k,t_ j)[1-F_ k(x_ k)][1-F_ j(x_ j)]\right\}, \] where \(F_ k\) is the marginal distribution function of \(X_{t_ k}\). He examines the restrictions on the values of the dependence coefficient \(\alpha\) and discusses their effect in detail for stationary processes. It is shown that the continuous time processes do not enjoy the usual regularity properties and the discrete time ones cannot have the well- known autoregressive and moving average correlation structures.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3973518)