Answers are grounded in this paper and cite their sources.
We consider the class A(φ,β) of random analytic functions in the unit disk C= |z|<1\ of the form f(z,ω)=f(z,ω₁,ω₂)=∑ₙ₌₀+∞ Rₙ(ω₁)ξₙ(ω₂)aₙzⁿ, where aₙ limn→+∞√[n]|aₙ|=1, (Rₙ(ω)) is the Rademacher sequence, (ξₙ(ω)) is a sequence of complex-valued random variables (denote by Δφ) such that there exists a constant β>0 and a function φ(N,β)₊→[1;+∞) non-decreasing by N and β for which $ (E(max0≤ n≤ N|ξ_n|β))1/βφ(N,β),\ \ N→+∞,\ \ α=N→+∞lnφ(N,β)/ln N<+∞,$ $ (∃ γ>0)(∃ n_0) \{ E|ξ_n|-γ\ n≥ n_0\}<+∞.$ By $A_1(φ,β)$ we denote the class of random analytic functions in $D$ of the form $f(z,ω)=∑ₙ₌₀+∞ ξ_n(ω)a_nz^n,$ where a sequence $(ξ_n(ω))∈Δ_φ$ and, in particular, may be not sub-gaussian and not independent. In the paper, there are proved the following statements: Let $δ>0.$ 1) Theorem 3: For $f(φ,β)$ there exist $r_0(ω)>0$, a set $E(δ)⊂(0;1)$ of finite logarithmic measure such that for all $r∈(r_0(ω);1) E$ we have with probability $p∈(0;1)$ $ M_f(r,ω) ≤{μ_f(r)}{(1-p)1/β}φ(N(r),β)((1-r)⁻²· lnμ_f(r)φ(N(r),β)/(1-p)(1-r))1/4+δ. $ 2) Theorem 4: For a function $f_1(φ,β)$ there exist $r_0(ω)>0$, a set $E(δ)⊂(0;1)$ of finite logarithmic measure such that for all $r∈(r_0(ω);1) E$ we get with probability $p∈(0;1)$ $ M_f(r,ω) ≤{μ_f(r)}{(1-p)1/β}φ(N(r),β)((1-r)⁻²· lnμ_f(r)φ(N(r),β)/(1-p)(1-r))1/2+δ. $
No takes yet. Share an insight, caveat, or question.
Kuryliak et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: