The weak law of large numbers for arrays (Q1198990): Difference between revisions

Let \(\{X_{ni},1\leq i\leq k_ n,\;n\geq 1\}\), \(k_ n\to\infty\) as \(n\to\infty\), be an array of random variables and \[ {1\over k_ n}\sum^{k_ n}_{i=1}E| X_{ni}|^ pI\{| X_{ni}|>a\}\to 0\quad\text{as }a\to\infty,\;0<p<2.\tag{*} \] Set \(S_ n=\sum^{k_ n}_{i=1}(X_{ni}-\mu_{ni})\) where \(\mu_{ni}=0\) if \(0<p<1\) and \(\mu_{ni}=E(X_{ni}\mid{\mathcal F}_{n,i- 1})\) with \({\mathcal F}_{n,i}=\sigma(X_{nj},1\leq j\leq i-1)\) for \(1\leq p<2\). It is proved that \(k_ n^{-1/p}S_ n\to 0\) in \(L^ p\) as \(n\to\infty\); see for \(p=1\) [\textit{T. K. Chandra}, Sankhyā, Ser A 51, No. 3, 309-317 (1989; Zbl 0721.60024)]. Together with \((*)\) some other domination conditions are listed which can serve by the extension of laws of large numbers from the case of i.i.d. random variables to the general one, relations between them are discussed as well as connections of the theorem proved with previous results.