Block empirical likelihood inference for stochastic bounding: large deviations asymptotics under m-dependence

The present note is occupied with the issue of generalized Neyman-Pearson optimality, for a testing procedure for the determination of stochastic bounding, that is based on data blocking and the minimization of the Kullback-Liebler divergence, in a time series context of m-dependence. Optimality is established via an extension of Sanov’s Theorem on empirical measures for blocks of data of temporal dependence that becomes asymptotically negligible at sufficiently fast rates. A large deviation property for the-subsequent to the derivation of the test statistic-BEL estimator, and a corresponding confidence region are also obtained.