Image transformations, Markov operators, and sample median

S. V. Butler · 2026

Read Image transformations, Markov operators, and sample median by S. V. Butler (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-03130.txt.

Introduction

S. V. Butler develops a broad theory of image transformations and Markov operators on spaces of topological measures, then applies it to the sample median. Iterated function systems and their operators are classic tools in fractal theory, and this work extends them to infinite mixtures.

Butler shows that when the transformations contract in a suitable metric, the operator has a unique invariant measure, recovering the classic fractal result as a special case. A generalized median distribution is defined and shown to behave well under a wide class of variable changes. A final result links low-dimensional topological measures to Radon measures.

Catalog record

CREATORS. V. Butler
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE91 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.03130

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Image transformations, Markov operators, and sample median S. V. Butler Abstract (I.) We consider generalizations of an iterated function system and the associated Markov operators. A Markov operator, defined on the space of (deficient) topological measures on a locally compact space, is an infinite convex linear combination of adjoints of (d-) image transformations. Restricted to measures, this Markov-Feller operator has a nonlinear dual operator given by an infinite convex linear combination of (conic) quasi-homomorphisms. If (d-) image transformations are contractions with respect to the Kantorovich-Rubinstein metric, a Markov operator has the unique invariant (deficient) topological measure. Taking a compact space, finitely many inverses of contractions as image transformations, and restricting the Markov operator to measures gives the classical result from the theory of fractals. There are various relations between Markov operator and the iterated function system where adjoints of (d-) image transformations are contractions on the compact metric space of { 0 , 1 } \{0,1\} -valued (deficient) topological measures. For instance, the invariant (deficient) topological measure is the composition of the fixed point of the IFS and the basic (d-) image transformation. (II.) We define a generalized distribution of the sample median (g.d.s.m.) for continuous proper maps using an image transformation. We show that the g.d.s.m. and the inverse on the sample median are equivariant under solid variables, a large collection of transformations. On ℝ n \mathbb{R}^{n} such transformations include rotations, translations, symmetries, stretching, projections, monotone maps, etc. (III.) We show that a (signed) topological measure on a locally compact space with the covering dimension dim X ≤ 1 \dim X\leq 1 is a (signed) Radon measure. MSC Classification]Primary: 28A80, 62G30, 60B05, 28A33, 47H99. Secondary: 60B10, 47H04, 47H07, 28C05, 28C15 keywords Image transformations, Markov operators, Image transformation systems and Iterated function systems, Kantorovich-Rubinstein metric, generalized distribution of the sample median pacs [ 1 Introduction This paper covers three topics. (I.) We consider generalizations of an iterated function system and the associated Markov operator. We discuss the unique fixed point and also obtain a Markov-Feller operator that has a nonlinear dual operator. (II.) We define a generalized distribution of the sample median (g.d.s.m.) for continuous proper maps and show that the g.d.s.m. and the inverse on the sample median are equivariant under a wide collection of transformations. Although our generalizations belong to very different areas – theory of fractals and probability \& statistics –they use the same tools, namely, certain quasi-linear maps, (d-) image transformations, and their adjoints. (III.) We show that a (signed) topological measure on a locally compact space with the covering dimension \dim X\leq 1 is a (signed) Radon measure. This paper relies on results from Butler:QLM , where we study (d-) image transformations and (conic) quasi-linear maps. Quasi-linear maps are nonlinear operators from C_{0}(X) (where X is a locally compact space) to a normed linear space E which are linear on the smallest subalgebras generated by single functions. i.e. on B(f)=\{\phi\circ f:\,\phi\in C(\overline{f(X)})\} (with \phi(0)=0 if X is noncompact) where f\in C_{0}(X) . ( f\in C_{0}(X) if for every \epsilon>0 the set \{x:|f|\geq\epsilon\} is compact.) Conic quasi-linear maps are nonlinear operators which preserve nonnegative linear combinations on positive cones generated by single functions, i.e. on A^{+}(f)=\{\phi\circ f:\ \phi\in C(\overline{f(X)}),\phi\mbox{ is nondecreasing}\} , (with \phi(0)=0 if X is noncompact) where f\in C_{0}(X) . When E=\mathbb{R} , we obtain quasi-linear functionals and p-conic quasi-linear functionals. Such nonlinear functionals (called quasi-integrals) correspond to topological measures and deficient topological measures, respectively. These set functions generalize measures, and their definitions are close to that of a regular Borel measure, but they also have some striking differences from Borel measures. For instance, (deficient) topological measures are not subadditive. The theory of quasi-linear functionals and topological measures, introduced by J. F. Aarnes Aarnes:TheFirstPaper , has origins in mathematical axiomatization and interpretations of quantum physics. This theory has connections with different areas in mathematics, including probability and statistics, fractals, and Choquet integrals, and has proved to be very influential in symplectic geometry, playing an important role in function theory on symplectic manifolds ( PoltRosenBook ). Quasi-linear maps from C_{0}(X) to C_{0}(Y) that behave like homomorphisms on singly generated subalgebras or cones are called (conic) quasi-homomorphisms. They correspond to (d-) image transformations, one of the main objects of this paper. Image transformations and d-image transformations generalize the idea of u^{-1} in image measures \mu\circ u^{-1} . They move subsets of one space to subsets of another space so that (deficient) topological measures on one space produce (deficient) topological measures on another space. The adjoint of a (d-) image transformation is a continuous operator between spaces of (deficient) topological measures equipped with the weak topology, and it plays an important role in this paper. We will give some background about (deficient) topological measures, quasi-integrals, (d-) image transformations and quasi-homomorphisms in the next section. More information about these concepts and interconnections between them can be found in Butler:QLM . We consider a generalization of an iterated function system and Markov operators involving (deficient) topological measures on a locally compact metric space. A well-known result of Hutchinson gives the existence of the invariant measure for operator M(\nu)=\sum_{i=1}^{n}\alpha_{i}\ \nu\circ u_{i}^{-1} associated with the iterated function system with probabilities (X,u_{i},\alpha_{i},i=1,\ldots,n) , see Hutchinson , ( Barnsley , Ch. IX) . Here u_{i} are contractions on a complete metric space X,\alpha_{i}>0,\sum_{i=1}^{n}\alpha_{i}=1 . Replacing finitely many u_{i}^{-1} by infinitely many (d-) image transformations leads to a (d-) image transformation system, a generalization of an iterated function system. We consider the associated continuous Markov operator S between spaces of (deficient) topological measures, where S(\mu)=\sum_{i=1}^{\infty}\alpha_{i}q_{i}^{*}(\mu) , an infinite convex linear combination of adjoints of (d-) image transformations. Restricted to the space of measures, such an operator is a Markov-Feller operator with a nonlinear dual operator T=\sum_{i=1}^{\infty}\alpha_{i}\theta_{i} given by an infinite convex linear combination of (conic) quasi-homomorphisms that correspond to (d-) image transformations. We show that any uniformly bounded in variation family \mathcal{M} of (deficient) topological measures is a complete metric space with respect to the Kantorovich-Rubinstein metric. We prove that the considered Markov operator on \mathcal{M} has a unique invariant (deficient) topological measure if (d-) image transformations are contractions with respect to the Kantorovich-Rubinstein metric. If we take a compact metric space as the underlying space X , finitely many image transformations, each of which is the inverse of a contraction on X , and restrict a Markov operator to measures on X , we obtain the classical result mentioned above. We consider various relations between the Markov operator and the iterated function system (IFS) where adjoints of (d-) image transformations are contractions on the compact metric space of \{0,1\} -valued (deficient) topological measures. For example, the invariant (deficient) topological measure is the composition of the fixed point of the IFS and the basic (d-) image transformation. These and other results are in Section 3 . Together with new results, we also generalize and expand many results of J. Aarnes, O. Johansen, A. Rustad, and O. Pedersen about finitely many image transformations and topological measures on compact spaces ( Aarnes:ITfirst , AarnesJohansenRustad , Pedersen ). In Section 4 we prove an important theorem linking dimension theory and topological measures. It says that a (signed) topological measure on a locally compact space with the covering

Read the complete text