Statistical analysis in multivariate sampling

The analysis of univariate count data involves pairs of random variables@ Yi; Xi of m observations, while in the bivariate and multivariate, analysis data of k pairs ofrandom variables Yij; Xij; i = I;:::; m; j = I;:::; k; are involved. The fundamental goal of the work is, based on the appropriate model assumptions to obtain good estimates for the attribute totals such as: estimating the defective or damage totals, ie, estimating the defective totals Y in a specific factory containing a known total amount of the productions, or estimating the proportions of those defective or damage totals with respect to the total amount of the items, as well as, constructing reliable confidence intervals for the proportion, or constructing confidence intervals for any linear combination of these proportions (which may describe some monetary gain or loss).