scientific article; zbMATH DE number 4054691

From MaRDI portal
Publication:3790384





Let \(\{X_{nk}\), \(1\leq k\leq n<\infty \}\) be a row-independent array of random variables, and define \(S_ n=\sum^{n}_{k=1}X_{nk}\) and \[ B_ n(u)=u^{-2}\sum^{n}_{k=1}E(X^{*2}_{nk} I(| X^*_{nk}| ou))\quad for\quad n\geq 1,\quad u>0, \] where \(X^*_{nk}\) is a symmetrized version of \(X_{nk}\). This paper presents certain bounds for the local probabilities \(P_ n(x,\lambda)=P(x\leq S_ n\leq x+\lambda)\). For example, it is shown that \[ \sup_{x}P_ n(x,\lambda)=O(b_ n^{-1})\text{ for every } \lambda \geq 0 \] if \(b_ n\to \infty\) satisfies \(B_ n(u)\geq c(b_ n/u)^{\nu}\) for some \(c>0\), \(\nu >0\) and all u in \([\Delta,b_ n]\), \(\Delta >0\). Conditions are also given under which constants \(c_ 1\) and \(c_ 2\) exist such that \[ c_ 1\leq \lambda_ n^{-1}b_ nP_ n(x,\lambda)\leq c_ 2 \] for some sequence \(\lambda_ n=o(b_ n)\).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3790384)