Generalization of the Nualart-Peccati criterion (Q282494)

Let \(\{X_n\}_{n\geq 1}\) be a sequence of multiple Wiener-Itô integrals of a fixed order \(p\geq 2\) such that \(\mathbb E[X_n^2]\rightarrow 1\) as \(n\rightarrow\infty\). The fourth moment theorem, due to \textit{D. Nualart} and \textit{G. Peccati} [Ann. Probab. 33, No. 1, 177--193 (2005; Zbl 1097.60007)] asserts that \(X_n\) converges in law to a standard normal distribution if and only if \(\mathbb E[X_n^4]\rightarrow 3\) as \(n\rightarrow\infty\). This result offers a drastic simplification of the method of moments in this setting. The main result of the present paper is a generalisation of this fourth moment theorem. The authors prove that, for \(\{X_n\}_{n\geq 1}\) as above and a fixed integer \(k\geq 2\), \(X_n\) converges in law to a standard normal distribution if and only if \(\mathbb E[X_n^{2k}]\rightarrow\mathbb E[N^{2k}]\), where \(N\sim\mathrm N(0,1)\). The authors go further and give an explicit rate of convergence in the total variation distance in this setting, in the spirit of [\textit{I. Nourdin} and \textit{G. Peccati}, Probab. Theory Relat. Fields 145, No. 1-2, 75--118 (2009; Zbl 1175.60053)]: \[ d_{TV}(X_n,\mathrm N(0,1))\leq C_k\sqrt{\frac{\mathbb E[X_n^{2k}]}{\mathbb E[N^{2k}]}-1}, \] where an explicit expression for the constant \(C_k\) is given. Several new inequalities for multiple Wiener-Itô integrals arise from the proofs, and may be of some independent interest. In fact, the above results on multiple Wiener-Itô integrals follow from more general and powerful results on eigenfunctions of a diffusive Markov operator. Finally, the paper concludes with two conjectures relevant to this framework, which would make interesting areas for future research.