Study of Power Series Distributions with Specified Covariances
Nataliia Voinalovych · 2025
Read Study of Power Series Distributions with Specified Covariances by Nataliia Voinalovych (2025) 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-2511-01081.txt.
Introduction
Oleksandr Volkov, Yurii Volkov, and Nataliia Voinalovych study power series distributions, a broad family that generalizes many familiar discrete laws, including the Bernoulli, Poisson, and geometric laws. Their aim is constructive: how to build such laws whose covariance structure is fixed in advance.
They derive a first order equation for the law's generating function. It ties its settings to the chosen covariance function. Choosing an analytic or polynomial covariance then fixes the generating function's structure. That link lets them construct new groups that keep a set covariance and extend the classical laws. The approach serves random modeling, where the spread pattern must be set ahead of time.
Catalog record
| CREATOR | Nataliia Voinalovych |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-01081
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Study of Power Series Distributions with Specified Covariances Nataliia Voinalovych Abstract This paper presents a study of power series distributions (PSD) with prescribed covariance characteristics. Such distributions constitute a fundamental class in probability theory and mathematical statistics, as they generalize a wide range of well-known discrete distributions and enable the description of various stochastic phenomena with a predetermined variance structure. The aim of the research is to develop analytical methods for constructing power series distributions with given covariances and to establish the conditions under which a particular function can serve as the covariance of a certain PSD. The paper derives a first-order differential equation for the generating function of the distribution, which determines the relationship between its parameters and the form of the covariance function. It is shown that the choice of an analytical or polynomial covariance completely specifies the structure of the corresponding generating function. Based on this, classes of distributions are established that realize linear, quadratic, and more general analytical dependencies between the expectation and the variance. The analysis made it possible to construct new families of PSDs that generalize the classical Bernoulli, Poisson, geometric, and other distributions while preserving a given covariance structure. The proposed approach is based on the analytical relationship between the generating function and the covariance function, providing a framework for constructing stochastic models with predefined dispersion properties. The results obtained expand the theoretical framework for describing discrete distributions and open up opportunities for practical applications in statistical estimation, modeling of complex systems, financial processes, machine learning where it is crucial to control the dependence between the mean and the variation. Further research may focus on constructing continuous analogues of such distributions, studying their limiting properties, and applying them to problems of regression and Bayesian analysis. с тудент 1-го року навчання магстерсько програми за спецальнстю <<Cтатистика>> Унверситету Калфорн, Берклстатистика, стохастичне моделювання, теоря операторв, машинне навчаннядоктор фз.-мат. наук, професортеоря ймоврностей, статистика, дискретна математикаканд. пед. наук, доцент кафедри математики, фзики та методик навчання Центральноукранського державного унверситету мен Володимира Винниченкаметодика навчання математики, теоря ймоврностей, статистикаUniversity of California at Berkeley, Master’s student at the Department of Statisticsstatistics, stochastic modeling, operator theory, machine learningDoctor of Physics an dMathematics, Professorprobability theory, statistics, discrete mathematicsCandidate of Pedagogical Sciences, Associate Professor of the Department of Mathematics and Digital Technologies, Volodymyr Vynnychenko Central Ukrainian State Universitymethodology of teaching mathematics, probability theory, statisticsДослдження розподлв степеневих рядв з наперед заданими коварацямиУ робот проведено дослдження розподлв степеневих рядв (РСР), для яких наперед задано коварацйн характеристики. Так розподли становлять фундаментальний клас у теор ймоврностей та математичнй статистиц, оскльки вони узагальнюють велику кльксть вдомих дискретних розподлв дозволяють описати широкий спектр стохастичних явищ з наперед визначеною структурою варац. Метою дослдження розробка аналтичних методв побудови розподлв степеневих рядв з заданими коварацями та встановлення умов, за яких певна функця може виступати коварацю деякого РСР. У статт отримано рвняння першого порядку для генератриси розподлу, що визнача взамозв’язок мж параметрами та формою коварацйно функц. Показано, що вибр аналтично або многочленно коварац повнстю визнача структуру вдповдно генератриси. На основ цього встановлено класи розподлв, як реалзують лнйн, квадратичн та бльш загальн аналтичн залежност мж математичним сподванням дисперсю. Проведений аналз дозволив побудувати нов смейства РСР, як узагальнюють класичн розподли Бернулл, Пуассона, геометричний та нш, збергаючи при цьому задану форму коварац. Запропонований пдхд базуться на аналтичному зв’язку мж генератрисою розподлу та функцю коварац забезпечу можливсть конструювання стохастичних моделей з наперед визначеними дисперсйними властивостями. Отриман результати розширюють теоретичний апарат опису дискретних розподлв вдкривають можливост для практичного застосування у статистичному оцнюванн, моделюванн складних систем, фнансових процесв, машинному навчанн, де важливо контролювати залежнсть мж середнм варацю. Подальш дослдження можуть бути спрямован на побудову неперервних аналогв таких розподлв, вивчення хнх граничних властивостей застосування у задачах регресйного та байсвського аналзу. Introduction Power series distributions (PSDs) were first systematically considered by Noack [ 1 ] in the context of generalizing classical discrete distributions. Since then, they have taken an important place in probability theory and stochastic modeling, particularly due to their ability to describe a wide class of random variables with specified generating function properties. Further development of the PSD theory is associated with the works of Johnson, Kotz, and Balakrishnan [ 2 ] , who presented a generalized theory of discrete distributions using power series in their fundamental work. Significant contributions to the study of the properties of such distributions were also made by Morgan [ 3 ] and Kemp [ 4 ] , who considered the classification and parameterization of PSDs based on their moment characteristics. Of particular interest are distributions for which the covariance characteristic (variance function) has a pre-specified form. Similar approaches have found application in statistical modeling where it is necessary to construct stochastic structures with predefined dependencies between the mean and the variance. In this context, the study of distributions with specified covariances is a natural continuation and generalization of the classical results obtained for the Poisson, Bernoulli, and geometric distributions [ 5 ] , [ 6 ] . In recent decades, there has been a growing interest in analytical methods for describing such distributions through equations for their generating functions, which allows for obtaining new classes of probability models with an analytical dependence between parameters [ 7 ] . In particular, Wolfowitz [ 8 ] developed a framework for analyzing distributions with specified moment characteristics, which is now also applied to PSDs. However, the question of constructing PSDs with a specified analytical form of the covariance remains open and is of both theoretical and applied interest. In particular, it is important to determine the conditions under which an arbitrary analytical or polynomial function can act as the covariance of a certain power series distribution. The relevance of the research is due to the necessity of expanding the class of known PSDs by introducing a parameterization based on specified covariances. This approach allows for constructing new types of probability distributions with controlled variance properties, which can be useful in problems of stochastic modeling, statistical estimation, and machine learning. Furthermore, establishing a connection between the functional form of the covariance and the corresponding generating function opens possibilities for further classification of PSDs and generalization of known results. The goal of the work is to investigate the properties of power series distributions with specified covariances, establish the equation satisfied by the generating function of such a distribution, and determine the conditions under which an arbitrary function can be the covariance of a PSD. Power Series Distributions with Specified Covariances The term ‘‘power series distributions’’ (PSD) belongs to Noack [ 1 ] . These distributions are defined as follows: let a power series with non-negative coefficients be given \omega(y)\mathchar 61\penalty\sum_{k\mathchar 61\penalty 0}^{\infty}a_{k}y^{k},\penalty\ a_{k}\geq\penalty 0,\penalty\ 0\leq\penalty y\mathchar 60\penalty R, ( R is the radius of convergence of the series). Consider a random variable \xi that can take non-negative integer values with probabilities p_{k}\mathchar 61\penalty P\{\xi\mathchar 61\penalty k\}:\mathchar 61\penalty a_{k}\frac{y^{k}}{\omega(y)},\penalty\ k\mathchar 61\penalty 0,1,2,\ldots Such a distribution is called the power series distribution of the function \omega(y) . Its probability generating function is P(z)\mathchar 61\penalty\sum_{k\mathchar 61\penalty 0}^{\infty}a_{k}\frac{y^{k}}{\omega(y)}z^{k}\mathchar 61\penalty\frac{\omega(yz)}{\omega(y)}. From this, we obtain the expected value and the variance of \xi . We have \text{M}\xi\mathchar 61\penalty P^{\prime}(1)\mathchar 61\penalty y\frac{\omega^{\prime}(y)}{\omega(y)}, \text{D}\xi\mathchar 61\penalty P^{\prime\prime}(1)\mathchar 43\penalty P^{\prime}(1)-(P^{\prime}(1))^{2}, and hence \text{D}\xi\mathchar 61\penalty y^{2}\frac{\omega^{\prime\prime}(y)}{\omega(y)}\mathchar 43\penalty y\frac{\omega^{\prime}(y)}{\omega(y)}-y^{2}\left(\frac{\omega^{\prime}(y)}{\omega(y)}\right)^{2}. Let’s introduce a new parameterization x\mathchar 61\penalty y\frac{\omega^{\prime}(y)}{\omega(y)}. (1) Then \text{D}\xi\mathchar 61\penalty y\frac{dx}{dy}.