Fourier coefficients of invariant random fields on homogeneous spaces of compact Lie groups (Q2346190): Difference between revisions
From MaRDI portal
Set profile property. |
Changed an Item |
||
Property / arXiv ID | |||
Property / arXiv ID: 1304.5142 / rank | |||
Normal rank |
Revision as of 04:29, 19 April 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Fourier coefficients of invariant random fields on homogeneous spaces of compact Lie groups |
scientific article |
Statements
Fourier coefficients of invariant random fields on homogeneous spaces of compact Lie groups (English)
0 references
29 May 2015
0 references
Let \(G\) be a compact group, \(K\) a closed subgroup of \(G\), and \(\mathfrak{X}\) the homogeneous space \(G/K\). Let \(V\subset L^2(\mathfrak{X})\) be a self-conjugated irreducible \(G\)-module, and \((v_i)_{-l\leq i\leq l}\) be a self-conjugated orthogonal basis of \(V\). Denote by \(D(g)\) the representative matrix of the action of \(G\) on \(V\). The authors say that the basis \((v_i)_{-l\leq i\leq l}\) is mixing if there is \(g\in G\) and \(0\leq m_1<m_2\leq l\) such that \(|D_{m_i, m}(g)|\neq |D_{m_i, -m}(g)|\) for every \(0<m\leq l\), \(i=1, 2\) (the mixing condition). It turns out that in some cases if the mixing condition holds, then the independence of the coefficients implies Gaussianity for an invariant random field. In particular, the following statement was proved by the author et al. [Electron. Commun. Probab. 12, 291--302 (2007; Zbl 1128.60039)]. Let \(G\) be a compact connected Lie group and \(T\) an a.s. square integrable real \(G\)-invariant random field on the homogeneous space \(\mathfrak{X}\) of \(G\). Let \(V\subset L^2(\mathfrak{X})\) be an irreducible \(G\)-module such that \(\overline V=V\). Assume that \((v_k)_{-l\leq k\leq l}\) is a self-conjugated mixing basis of \(V\). Consider the real random field \(T_V(x)+\sum_k a_kv_k(x), \) where the r.v.'s \(a_k\), \(k\geq 0\) are independent. Then, if \(T_V\) is \(G\)-invariant, the r.v.'s \(a_k\) are jointly Gaussian and therefore also \(T_V\) is Gaussian. The authors investigate when the mixing condition is valid or not and obtain some new interesting results similar to the above mentioned one.
0 references
random fields
0 references
Fourier expansions
0 references
characterization of Gaussian random fields
0 references