Showing 1–21 of 21 results for author: Basor, E L

Diagonal Nonlinear Transformations Preserve Structure in Covariance and Precision Matrices

Authors: Rebecca E Morrison, Ricardo Baptista, Estelle L Basor

Abstract: For a multivariate normal distribution, the sparsity of the covariance and precision matrices encodes complete information about independence and conditional independence properties. For general distributions, the covariance and precision matrices reveal correlations and so-called partial correlations between variables, but these do not, in general, have any correspondence with respect to independ… ▽ More For a multivariate normal distribution, the sparsity of the covariance and precision matrices encodes complete information about independence and conditional independence properties. For general distributions, the covariance and precision matrices reveal correlations and so-called partial correlations between variables, but these do not, in general, have any correspondence with respect to independence properties. In this paper, we prove that, for a certain class of non-Gaussian distributions, these correspondences still hold, exactly for the covariance and approximately for the precision. The distributions -- sometimes referred to as "nonparanormal" -- are given by diagonal transformations of multivariate normal random variables. We provide several analytic and numerical examples illustrating these results. △ Less

Submitted 20 September, 2021; v1 submitted 8 July, 2021; originally announced July 2021.

Asymptotic formulas for determinants of a special class of Toeplitz + Hankel matrices

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: We compute the asymptotics of the determinants of certain $n\times n$ Toeplitz + Hankel matrices $T_n(a)+H_n(b)$ as $n\to\infty$ with symbols of Fisher-Hartwig type. More specifically we consider the case where $a$ has zeros and poles and where $b$ is related to $a$ in specific ways. Previous results of Deift, Its and Krasovsky dealt with the case where $a$ is even. We are generalizing this in a m… ▽ More We compute the asymptotics of the determinants of certain $n\times n$ Toeplitz + Hankel matrices $T_n(a)+H_n(b)$ as $n\to\infty$ with symbols of Fisher-Hartwig type. More specifically we consider the case where $a$ has zeros and poles and where $b$ is related to $a$ in specific ways. Previous results of Deift, Its and Krasovsky dealt with the case where $a$ is even. We are generalizing this in a mild way to certain non-even symbols. △ Less

Submitted 1 March, 2016; originally announced March 2016.

MSC Class: 47B35

Asymptotics of determinants of Hankel matrices via non-linear difference equations

Authors: Estelle L. Basor, Yang Chen, Nazmus S. Haq

Abstract: E. Heine in the 19th century studied a system of orthogonal polynomials associated with the weight $\left[x(x-α)(x-β)\right]^{-\frac{1}{2}}$, $x\in[0,α]$, $0<α<β$. A related system was studied by C. J. Rees in 1945, associated with the weight $\left[(1-x^2)(1-k^2x^2)\right]^{-\frac{1}{2}}$, $x\in[-1,1]$, $k^2\in(0,1)$. These are also known as elliptic orthogonal polynomials, since the moments of t… ▽ More E. Heine in the 19th century studied a system of orthogonal polynomials associated with the weight $\left[x(x-α)(x-β)\right]^{-\frac{1}{2}}$, $x\in[0,α]$, $0<α<β$. A related system was studied by C. J. Rees in 1945, associated with the weight $\left[(1-x^2)(1-k^2x^2)\right]^{-\frac{1}{2}}$, $x\in[-1,1]$, $k^2\in(0,1)$. These are also known as elliptic orthogonal polynomials, since the moments of the weights maybe expressed in terms of elliptic integrals. Such orthogonal polynomials are of great interest because the corresponding Hankel determinant, depending on a parameter $k^2$, where $0<k^2<1$ is the $τ$~function of a particular Painlevé VI, the special cases of which are related to enumerative problems arising from string theory. We show that the recurrence coefficients, denoted by $β_n(k^2),\;n=1,2,\dots$; and ${\rm p}_1(n,k^2)$, the coefficients of $x^{n-2}$ of the monic polynomials orthogonal with respect to a generalized version of the weight studied by Rees, $$ (1-x^2)^α(1-k^2x^2)^β,\;\;x\in[-1,1],\;α>-1,\;β\in \mathbb{R}, $$ satisfy second order non-linear difference equations. The large $n$ expansion based on the difference equations when combined with known asymptotics of the leading terms of the associated Hankel determinant yields a complete asymptotic expansion of the Hankel determinant. The Painlevé equation is also discussed as well as the generalization of the linear second order differential equation found by Rees. △ Less

Submitted 15 June, 2015; v1 submitted 9 January, 2014; originally announced January 2014.

Comments: 47 pages, 1 table. V2: 49 pages, 1 table, Minor typos and formatting issues corrected

MSC Class: 15B52; 33C47; 34E05; 42C05; 47B35

Journal ref: Journal of Approximation Theory, 198 (2015) 63-110

Fredholm and invertibility theory for a special class of Toeplitz + Hankel operators

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: We develop a complete Fredholm and invertibility theory for Toeplitz+Hankel operators $T(a)+H(b)$ on the Hardy space $H^p$, $1<p<\infty$, with piecewise continuous functions $a,b$ defined on the unit circle which are subject to the condition $a(t)a(t^{-1})=b(t)b(t^{-1})$, $|t|=1$. In particular, in the case of Fredholmness, formulas for the defect numbers are established. The results are applied t… ▽ More We develop a complete Fredholm and invertibility theory for Toeplitz+Hankel operators $T(a)+H(b)$ on the Hardy space $H^p$, $1<p<\infty$, with piecewise continuous functions $a,b$ defined on the unit circle which are subject to the condition $a(t)a(t^{-1})=b(t)b(t^{-1})$, $|t|=1$. In particular, in the case of Fredholmness, formulas for the defect numbers are established. The results are applied to several important examples. △ Less

Submitted 4 October, 2011; originally announced October 2011.

Comments: 38 pages, 6 figures

MSC Class: 47B35

Determinant computations for some classes of Toeplitz-Hankel matrices

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: The purpose of this paper is to compute the asymptotics of determinants of finite sections of operators that are trace class perturbations of Toeplitz operators. For example, we consider the asymptotics in the case where the matrices are of the form $ (a_{i-j} \pm a_{i+j+1-k})_{i,j=0... N-1} $ with $k$ is fixed. We will show that this example as well as some general classes of operators have exp… ▽ More The purpose of this paper is to compute the asymptotics of determinants of finite sections of operators that are trace class perturbations of Toeplitz operators. For example, we consider the asymptotics in the case where the matrices are of the form $ (a_{i-j} \pm a_{i+j+1-k})_{i,j=0... N-1} $ with $k$ is fixed. We will show that this example as well as some general classes of operators have expansions that are similar to those that appear in the Strong Szegö Limit Theorem. We also obtain exact identitities for some of the determinants that are analogous to the one derived independently by Geronimo and Case and by Borodin and Okounkov for finite Toeplitz matrices. These problems were motivated by considering certain statistical quantities that appear in random matrix theory. △ Less

Submitted 18 April, 2008; originally announced April 2008.

Report number: AIM 2008-23 MSC Class: 47B35

Asymptotics of block Toeplitz determinants and the classical dimer model

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: We compute the asymptotics of a block Toeplitz determinant which arises in the classical dimer model for the triangular lattice when considering the monomer-monomer correlation function. The model depends on a parameter interpolating between the square lattice ($t=0$) and the triangular lattice ($t=1$), and we obtain the asymptotics for $0<t\le 1$. For $0<t<1$ we apply the Szegö Limit Theorem fo… ▽ More We compute the asymptotics of a block Toeplitz determinant which arises in the classical dimer model for the triangular lattice when considering the monomer-monomer correlation function. The model depends on a parameter interpolating between the square lattice ($t=0$) and the triangular lattice ($t=1$), and we obtain the asymptotics for $0<t\le 1$. For $0<t<1$ we apply the Szegö Limit Theorem for block Toeplitz determinants. The main difficulty is to evaluate the constant term in the asymptotics, which is generally given only in a rather abstract form. △ Less

Submitted 28 July, 2006; originally announced July 2006.

MSC Class: 47B35 82B05

Extremal Non-Compactness of Composition Operators with Linear Fractional Symbol

Authors: Estelle L. Basor, Dylan Q. Retsek

Abstract: We realize norms of most composition operators acting on the Hardy space with linear fractional symbol as roots of hypergeometric functions. This realization leads to simple necessary and sufficient conditions on the symbol to exhibit extremal non-compactness, establishes equivalence of cohyponormality and cosubnormality of composition operators with linear fractional symbol, and yields a comple… ▽ More We realize norms of most composition operators acting on the Hardy space with linear fractional symbol as roots of hypergeometric functions. This realization leads to simple necessary and sufficient conditions on the symbol to exhibit extremal non-compactness, establishes equivalence of cohyponormality and cosubnormality of composition operators with linear fractional symbol, and yields a complete classification of those linear fractional that induce composition operators whose norms are determined by the action of the adjoint on the normalized reproducing kernels in the Hardy space. △ Less

Submitted 25 July, 2005; originally announced July 2005.

MSC Class: 47B33

On the asymptotics of certain Wiener-Hopf-plus-Hankel determinants

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: In this paper we determine the asymptotics of the determinants of truncated Wiener-Hopf plus Hankel operators $\det(W_R(a)\pm H_R(a))$ as $R$ tends to infinity for symbols $a(x)=(x^2/(1+x^2))^β$ with the parameter $β$ being of small size. In this paper we determine the asymptotics of the determinants of truncated Wiener-Hopf plus Hankel operators $\det(W_R(a)\pm H_R(a))$ as $R$ tends to infinity for symbols $a(x)=(x^2/(1+x^2))^β$ with the parameter $β$ being of small size. △ Less

Submitted 1 February, 2005; originally announced February 2005.

Comments: 38 pages

MSC Class: 47B35

Factorization of a class of Toeplitz + hankel operators and the A_p-condition

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: Let $M(φ)=T(φ)+H(φ)$ be the Toeplitz plus Hankel operator acting on $H^p(\T)$ with generating function $φ\in L^\iy(\T)$. In a previous paper we proved that $M(φ)$ is invertible if and only if $φ$ admits a factorization $φ(t)=φ_{-}(t)φ_{0}(t)$ such that $φ_{-}$ and $φ_{0}$ and their inverses belong to certain function spaces and such that a further condition formulated in terms of $φ_{-}$ and… ▽ More Let $M(φ)=T(φ)+H(φ)$ be the Toeplitz plus Hankel operator acting on $H^p(\T)$ with generating function $φ\in L^\iy(\T)$. In a previous paper we proved that $M(φ)$ is invertible if and only if $φ$ admits a factorization $φ(t)=φ_{-}(t)φ_{0}(t)$ such that $φ_{-}$ and $φ_{0}$ and their inverses belong to certain function spaces and such that a further condition formulated in terms of $φ_{-}$ and $φ_{0}$ is satisfied. In this paper we prove that this additional condition is equivalent to the Hunt-Muckenhoupt-Wheeden condition (or, $A_{p}$-condition) for a certain function $σ$ defined on $[-1,1]$, which is given in terms of $φ_{0}$. As an application, a necessary and sufficient criteria for the invertibility of $M(φ)$ with piecewise continuous functions $φ$ is proved directly. Fredholm criteria are obtained as well. △ Less

Submitted 20 April, 2004; originally announced April 2004.

MSC Class: 47B35: 47G

On the Determinant of a Certain Wiener-Hopf + Hankel Operator

Authors: Estelle L. Basor, Torsten Ehrhardt, Harold Widom

Abstract: We establish an asymptotic formula for determinants of truncated Wiener-Hopf+Hankel operators with symbol equal to the exponential of a constant times the characteristic function of an interval. This is done by reducing it to the corresponding (known) asymptotics for truncated Toeplitz+Hankel operators. The determinants in question arise in random matrix theory in determining the limiting distri… ▽ More We establish an asymptotic formula for determinants of truncated Wiener-Hopf+Hankel operators with symbol equal to the exponential of a constant times the characteristic function of an interval. This is done by reducing it to the corresponding (known) asymptotics for truncated Toeplitz+Hankel operators. The determinants in question arise in random matrix theory in determining the limiting distribution for the number of eigenvalues in an interval for a scaled Laguerre ensemble of positive Hermitian matrices. △ Less

Submitted 31 March, 2003; originally announced April 2003.

Comments: LaTeX file, 14 pages

MSC Class: 47B35

The X-ray problem revisited

Authors: Estelle L. Basor, Yang Chen

Abstract: In this letter we re-visit the X-ray problem. Assuming point interaction between the conduction electrons and the first instantaneously created core-hole, the latter's Green's function can be represented as a Fredholm determinant of certain Wiener-Hopf operators acting on L^2(0,T) with discontinuous symbols. Here the symbols are the local conduction electron Green's function in the frequency dom… ▽ More In this letter we re-visit the X-ray problem. Assuming point interaction between the conduction electrons and the first instantaneously created core-hole, the latter's Green's function can be represented as a Fredholm determinant of certain Wiener-Hopf operators acting on L^2(0,T) with discontinuous symbols. Here the symbols are the local conduction electron Green's function in the frequency domain and T is the time the core-hole spends in the system before removal. In this situation, the classical theory of singular integral equations usually employed in the literature to compute the large T asymptotics of the Fredholm determinant ceased to be applicable. A rigorous theory first put forward in the context of operator theory comes into play and universal constants are found in the aymptotics. △ Less

Submitted 12 February, 2003; originally announced February 2003.

Comments: 8 pages 0 figures

Wiener-Hopf determinants with Fisher-Hartwig symbols

Authors: Estelle L. Basor, Harold Widom

Abstract: With localization techniques one can obtain general limit theorems for Toeplitz determinants with Fisher-Hartwig singularities from the asymptotics for any symbol with one singularity of general type. There exists a family of these for which the determinants can be evaluated explicitly and their asymptotics determined. But for the Wiener-Hopf analogue, although there are likely analogous localiz… ▽ More With localization techniques one can obtain general limit theorems for Toeplitz determinants with Fisher-Hartwig singularities from the asymptotics for any symbol with one singularity of general type. There exists a family of these for which the determinants can be evaluated explicitly and their asymptotics determined. But for the Wiener-Hopf analogue, although there are likely analogous localization techniques, there is not a single example known of a symbol with Fisher-Hartwig singularity for which the determinant can be evaluated explicitly. In this paper we determine the asymptotics of Wiener-Hopf determinants for a symbol with one Fisher-Hartwig singularity of general type. We do this by showing that it is asymptotically equal to a Toeplitz determinant with symbol having the corresponding singularity. △ Less

Submitted 9 June, 2003; v1 submitted 17 May, 2002; originally announced May 2002.

MSC Class: 47B35; 82B

Asymptotics of determinants of Bessel operators

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: In this paper we determine the asymptotics of the determinant of Bessel operators for sufficiently smooth generating functions. These operators are similar to Wiener-Hopf operators with the Fourier transform replaced by the Hankel transform and thus the asymptotics of the determinanst are similar to the well-known Szegö-Akhiezer-Kac formula for truncated Wiener-Hopf determinants. In order to com… ▽ More In this paper we determine the asymptotics of the determinant of Bessel operators for sufficiently smooth generating functions. These operators are similar to Wiener-Hopf operators with the Fourier transform replaced by the Hankel transform and thus the asymptotics of the determinanst are similar to the well-known Szegö-Akhiezer-Kac formula for truncated Wiener-Hopf determinants. In order to compute the above, we also show that the Bessel operators differ from the Wiener-Hopf by a Hilbert-Schmidt operator. △ Less

Submitted 24 April, 2002; originally announced April 2002.

MSC Class: 47B35; 82B

Asymptotic formulas for the determinants of symmetric Toeplitz + Hankel matrices

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: We establish asymptotic formulas for the determinants of finite Toeplitz + Hankel matrices of size N, as N goes to infinity for singular generating functions defined on the unit circle in the special case where the generating function is even, i.e., where the Toeplitz + Hankel matrices are symmetric. We establish asymptotic formulas for the determinants of finite Toeplitz + Hankel matrices of size N, as N goes to infinity for singular generating functions defined on the unit circle in the special case where the generating function is even, i.e., where the Toeplitz + Hankel matrices are symmetric. △ Less

Submitted 8 February, 2002; originally announced February 2002.

MSC Class: 47B35

Distribution Functions for Random Variables for Ensembles of positive Hermitian Matrices

Authors: Estelle L. Basor

Abstract: Distribution functions for random variables that depend on a parameter are computed asymptotically for ensembles of positive Hermitian matrices. The inverse Fourier transform of the distribution is shown to be a Fredholm determinant of a certain operator that is an analogue of a Wiener-Hopf operator. The asymptotic formula shows that up to the terms of order $o(1)$, the distributions are Gaussia… ▽ More Distribution functions for random variables that depend on a parameter are computed asymptotically for ensembles of positive Hermitian matrices. The inverse Fourier transform of the distribution is shown to be a Fredholm determinant of a certain operator that is an analogue of a Wiener-Hopf operator. The asymptotic formula shows that up to the terms of order $o(1)$, the distributions are Gaussian. △ Less

Submitted 20 July, 2001; originally announced July 2001.

MSC Class: 47A75 82B

Journal ref: Communications in Mathematical Physics 188, 327-350

Some identities for determinants of structured matrices

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: In this paper we establish several relations between the determinants of the following structured matrices: Hankel matrices, symmetric Toeplitz + Hankel matrices and Toeplitz matrices. Using known results for the asymptotic behavior of Toeplitz determinants, these identities are used in order to obtain Fisher-Hartwig type results on the asymptotics of certain skewsymmetric Toeplitz determinants… ▽ More In this paper we establish several relations between the determinants of the following structured matrices: Hankel matrices, symmetric Toeplitz + Hankel matrices and Toeplitz matrices. Using known results for the asymptotic behavior of Toeplitz determinants, these identities are used in order to obtain Fisher-Hartwig type results on the asymptotics of certain skewsymmetric Toeplitz determinants and certain Hankel determinants. △ Less

Submitted 9 August, 2000; originally announced August 2000.

Determinants of Hankel Matrices

Authors: Estelle L. Basor, Yang Chen, Harold Widom

Abstract: The purpose of this paper is to compute asymptotically Hankel determinants for weights that are supported in a semi-infinite interval.The main idea is to reduce the problem to determinants of other operators whose determinant asymptotics are well known. The purpose of this paper is to compute asymptotically Hankel determinants for weights that are supported in a semi-infinite interval.The main idea is to reduce the problem to determinants of other operators whose determinant asymptotics are well known. △ Less

Submitted 8 June, 2000; originally announced June 2000.

Comments: 18 pages, LaTeX file

On a Toeplitz determinant identity of Borodin and Okounkov

Authors: Estelle L. Basor, Harold Widom

Abstract: The authors of the title proved an elegant identity expressing a Toeplitz determinant in terms of the Fredholm determinant of an infinite matrix which (although not described as such) is the product of two Hankel matrices. The proof used combinatorial theory, in particular a theorem of Gessel expressing a Toeplitz determinant as a sum over partitions of products of Schur functions. The purpose o… ▽ More The authors of the title proved an elegant identity expressing a Toeplitz determinant in terms of the Fredholm determinant of an infinite matrix which (although not described as such) is the product of two Hankel matrices. The proof used combinatorial theory, in particular a theorem of Gessel expressing a Toeplitz determinant as a sum over partitions of products of Schur functions. The purpose of this note is to give two other proofs of the identity. The first uses an identity of the second author for the quotient of Toeplitz determinants in which the same product of Hankel matrices appears and the second, which is more direct and extends the identity to the case of block Toeplitz determinants, consists of carrying the first author's collaborative proof of the strong Szegö limit theorem one step further. △ Less

Submitted 9 April, 2000; v1 submitted 1 September, 1999; originally announced September 1999.

Comments: 5 pages, LaTeX file. Misprints corrected

Determinants of Airy Operators and Applications to Random Matrices

Authors: Estelle L. Basor, Harold Widom

Abstract: The purpose of this paper is to describe asymptotic formulas for determinants of certain operators that are analogues of Wiener-Hopf operators. The determinant formulas yield information about the distribution functions for certain random variables that arise in random matrix theory when one rescales at the edge of the spectrum. The purpose of this paper is to describe asymptotic formulas for determinants of certain operators that are analogues of Wiener-Hopf operators. The determinant formulas yield information about the distribution functions for certain random variables that arise in random matrix theory when one rescales at the edge of the spectrum. △ Less

Submitted 13 October, 1999; v1 submitted 7 December, 1998; originally announced December 1998.

Comments: LaTeX file. 17 pages. An error corrected

Journal ref: J. Stat. Phys. 96 (1999) 1-20

Asymptotic formulas for determinants of a sum of finite Toeplitz and Hankel matrices

Authors: Estelle L. Basor, Torsten Ehrhardt

Abstract: The purpose of this paper is to describe asymptotic formulas for determinants of a sum of finite Toeplitz and Hankel matrices with singular generating functions. The formulas are similar to those of the analogous problem for finite Toeplitz matrices for a certain class of symbols. However, the appearance of the Hankel matrices changes the nature of the asymptotics in some instances depending on… ▽ More The purpose of this paper is to describe asymptotic formulas for determinants of a sum of finite Toeplitz and Hankel matrices with singular generating functions. The formulas are similar to those of the analogous problem for finite Toeplitz matrices for a certain class of symbols. However, the appearance of the Hankel matrices changes the nature of the asymptotics in some instances depending on the location of the singularities. Several concrete examples are also described in the paper. △ Less

Submitted 16 September, 1998; originally announced September 1998.

MSC Class: 47B35; 82B

Variance Calculations and the Bessel Kernel

Authors: Estelle L. Basor, Craig A. Tracy

Abstract: In the Laguerre ensemble of N x N (positive) hermitian matrices, it is of interest both theoretically and for applications to quantum transport problems to compute the variance of a linear statistic, denoted var_N f, as N->infinity. Furthermore, this statistic often contains an additional parameter alpha for which the limit alpha->infinity is most interesting and most difficult to compute numeri… ▽ More In the Laguerre ensemble of N x N (positive) hermitian matrices, it is of interest both theoretically and for applications to quantum transport problems to compute the variance of a linear statistic, denoted var_N f, as N->infinity. Furthermore, this statistic often contains an additional parameter alpha for which the limit alpha->infinity is most interesting and most difficult to compute numerically. We derive exact expressions for both lim_{N->infinity} var_N f and lim_{alpha->infinity}lim_{N->infinity} var_N f. △ Less

Submitted 1 February, 1999; v1 submitted 16 April, 1993; originally announced April 1993.

Comments: 7 pages; resubmitted to make postscript compatible

Journal ref: J. Stat. Phys. 73 (1993) 415-421