A Central Limit Theorem for the Sum of Generalized Linear and Quadratic Forms (Q4943302)

Let \(X_1,\dots,X_n\) be independent random variables and \(W_{kj}= W_{kjn} (X_k,X_j)\), \(L_k=L_{kn}(X_k)\) associated transformed variables with \(EL_k=0\), \(W_{kj}=W_{jk}\), \(E(W_{kj}\mid X_k)=0\) a.s. Consider \(T(n)=W(n)+ L(n)\), where \(W(n)= \sum\sum_{1\leq j<k\leq n}W_{kj}\) and \(L(n)=\sum^n_{j=1}L_j\). The aim of the paper is to give sufficient conditions for asymptotic normality of \(T(n)\). In particular, it is shown that a mild correlation type condition is sufficient to ensure that Lyapunov type conditions on the linear \(L(n)\) and quadratic \(W(n)\) forms individually will imply asymptotic normality of their sum \(T(n)\). The theorem generalizes CLT results of \textit{P. de Jong} [Probab. Theory Relat. Fields 75, 261-277 (1987; Zbl 0596.60022)] obtained for generalized quadratic forms \(W(n)\).