arXiv subjects

Wayne M. Lawton

Publications and source records attributed to Wayne M. Lawton.

6 recordsLinked to original sources

arXiv2023

A Mathematical Analysis of Benford's Law and its Generalization

We explain Kossovsky's generalization of Benford's law which is a formula that approximates the distribution of leftmost digits in finite sequences of natural data and apply it to six sequences of data including populations of US cities and towns and times between earthquakes. We model the natural logarithms of these two data sequences as samples of random variables having normal and reflected Gumbel densities respectively. We show that compliance with the general law depends on how nearly constant the periodized density functions are and that the models are generally more compliant than the natural data. This surprising result suggests that the generalized law might be used to improve density estimation which is the basis of statistical pattern recognition, machine learning and data science.

stat.ME↗

arXiv2023

Multidimensional Fourier Quasicrystals I. Sufficient Conditions

We derive sufficient conditions for an atomic measure $\sum_{λ\in Λ} m_λ\, δ_λ,$ where $Λ\subset \mathbb R^n,$ $m_λ$ are positive integers, and $δ_λ$ is the point measure at $λ,$ to be a Fourier quasicrystal, and suggest why they may also be necessary. These conditions extend the necessary and sufficient conditions derived by Lev, Olevskii, and Ulanovskii for $n = 1.$ Our methods exploit the toric geometry relation between Grothendieck residues and Newton polytopes derived by Gelfond and Khovanskii.

math.AG↗

arXiv2021

Bohr Almost Periodic Sets of Toral Type

A locally finite multiset $(Λ,c),$ $Λ\subset \mathbb R^n, c : Λ\rightarrow \{1,...,b\}$ defines a Radon measure $μ:= \sum_{λ\in Λ} c(λ)\, δ_λ$ that is Bohr almost periodic in the sense of Favorov if the convolution $μ*f$ is Bohr almost periodic every $f \in C_c(\mathbb R^n).$ If it is of toral type: the Fourier transform $\mathfrak F μ$ equals zero outside of a rank $m < \infty$ subgroup, then there exists a compactification $ψ: \mathbb R^n \rightarrow \mathbb T^m$ of $\mathbb R^n,$ a foliation of $\mathbb T^m,$ and a pair $(K,κ)$ where $K := \overline {ψ(Λ)}$ and $κ$ is a measure supported on $K$ such that $\mathfrak F κ= (\mathfrak F μ) \circ \widehat ψ$ where $\widehat ψ: \widehat {\mathbb T^m} \rightarrow \widehat {\mathbb R^n}$ is the Pontryagin dual of $ψ.$ If $(Λ,c)$ is uniformly discrete Bohr almost periodic and $c = 1,$ we prove that every connected component of $K$ is homeomorphic to $\mathbb T^{m-n}$ embedded transverse to the foliation and the homotopy of its embedding is a rank $m-n$ subgroup $S$ of $\mathbb Z^m,$ and we compute the density of $Λ$ as a function of $ψ$ and the homotopy of comonents of $K.$ For $n = 1$ and $K$ a nonsingular real algebraic variety, this construction gives all Fourier quasicrystals (FQ) recently characterized by Olevskii and Ulanovskii and suggest how to characterize FQ for $n > 1.$

math.CA↗

arXiv2021

Note on Spectral Factorization Results of Krein and Levin

Bohr proved that a uniformly almost periodic function $f$ has a bounded spectrum if and only if it extends to an entire function $F$ of exponential type $τ(F) < \infty$. If $f \geq 0$ then a result of Krein implies that $f$ admits a factorization $f = |s|^2$ where $s$ extends to an entire function $S$ of exponential type $τ(S) = τ(F)/2$ having no zeros in the open upper half plane. The spectral factor $s$ is unique up to a multiplicative factor having modulus $1.$ Krein and Levin constructed $f$ such that $s$ is not uniformly almost periodic and proved that if $f \geq m > 0$ has absolutely converging Fourier series then $s$ is uniformly almost periodic and has absolutely converging Fourier series. We derive neccesary and sufficient conditions on $f \geq m > 0$ for $s$ to be uniformly almost periodic, we construct an $f \geq m > 0$ with non absolutely converging Fourier series such that $s$ is uniformly almost periodic, and we suggest research questions.

math.CA↗

arXiv2014

Filters and Functions in Multi-scale Constructions: Extended Abstract

We derive results about geometric means of the Fourier modulus of filters and functions related to refinable distributions with arbitrary dilations and translations. Then we develop multi-scale constructions for dilations by Pisot-Vijayaraghavan numbers and translations in associated quasilattices.

math.CA↗

arXiv2011

Bose and Einstein Meet Newton

We model the time evolution of a Bose-Einstein condensate, subject to a special periodically excited optical lattice, by a unitary quantum operator U on a Hilbert space H. If a certain parameter alpha = p/q, where p and q are coprime positive integers, then H = L^2(R/Z,C^q) and U is represented by a q x q matrix-valued function M on R/Z that acts pointwise on functions in H. The dynamics of the quantum system is described by the eigenvalues of M. Numerical computations show that the characteristic polynomial det(zI - M(t)) = Prod_j=1^q (z - lambda_j(t)) where each lambda_j is a real analytic functions that has period 1/q. We discuss this phenomena using Newton's Theorem, published in Geometria analytica in 1660, and modern concepts from analytic geometry.

math.AG↗