On properties of sub-independent random variables (Q1080260)

Many of the convergence theorems in mathematical statistics as well as the theorems characterizing the normal distribution have, as part of their hypotheses, the requirement that two or more random variables be independent. In this work we use a concept of sub-independence, based only on the distribution of the sum, to prove a number of theorems related to the normal distribution. This same concept was used by the authors in Arch. Math. 43, 258-264 (1984; Zbl 0553.60028) to prove a version of the central limit theorem.