NEW TAIL PROBABILITY TYPE INEQUALITIES AND COMPLETE CONVERGENCE FOR WOD RANDOM VARIABLES WITH APPLICATION TO LINEAR MODEL GENERATED BY WOD ERRORS

Abstract

In this paper, we prove an exponential type inequality for unbounded Widely Orthant Dependent (WOD) random variables. Then we will use this inequality for establishing the almost complete convergence for a sequence of widely dependent random variables (WOD). As an application, we combine Markov inequalities and the new exponential inequality to discuss almost complete convergence of errors generated by a sequence of dependent random variables (WOD) in linear models.