Almost sure convergence and complete convergence for the weighted sums of martingale differences (Q4500099)

Consider a sequence \(\{(D_n,{\mathcal F}_n) :n\geq 1\}\) of martingale differences together with an array \(\{a_{ni}:1\leq i\leq n,n\geq 1\}\) of possible weights. On assuming that the sequence \(\{D_n:n\geq 1\}\) is stochastically bounded by some random variable \(D_0\) and that \(D_0\) satisfies either a suitable moment condition or a certain tail behaviour, the authors prove two strong laws of large numbers for weighted sums \(S_n=\sum^n_{i=1}a_{ni}X_i\), in which the weights \(a_{ni}\) are related to the tail behaviour of \(D_0\) in a particular way. Under similar assumptions, complete convergence is established for partial sums of Banach space valued random variables based on moving averages of a martingale difference sequence. The results extend earlier strong laws or Marcinkiewicz-Zygmund type laws for independent random variables.