Searcharxiv⌕ Search

arXiv subjects

Leo Egghe

Publications and source records attributed to Leo Egghe.

16 recordsLinked to original sources
arXiv2025

Mathematical reflections on modified fractional counting

We make precise what is meant by stating that modified fractional counting (MFC) lies between full counting and complete-normalized fractional counting by proving that for individuals, the MFC-values are weighted geometric averages of these two extremes. There are two essentially different ways to consider the production of institutes in multi-institutional articles, namely participation and actual number of contributions. Starting from an idea published by Sivertsen, Rousseau and Zhang in 2019 we present three formulae for measuring the production of institutes in multi-institutional articles. It is shown that the one proposed by Sivertsen, Rousseau and Zhang is situated between the two other ways. Less obvious properties of MFC are proven using the majorization order.

math.HO↗
arXiv2025

The Lorenz order in graph theory: A new proof and extension of the theorems of Hakimi and of Havel-Hakimi

This paper studies the relation between the Lorenz majorization order and the realizability of degree sequences X of a network in the sense of being graphical or connected graphical (c-graphical) or not. We prove the main result that, if X is dominated (in the Lorenz majorization sense) by X' and X' is (c-) graphical, the X is also (c-) graphical. We present a simple proof and a generalization of the Havel-Hakimi theorem, using the Lorenz order formalism. From this, a classical result of Hakimi on trees follows but also a new generalization to general connected networks. From this, a characterization of c-graphical sequences in terms of the Lorenz majorization order is given.

math.GM↗
arXiv2024

Spider networks

In this investigation we study a family of networks, called spiders, which covers a range of networks going from chains to complete graphs. These spiders are characterized by three parameters: the number of nodes in the core, the number of legs at each core node, and the length of these legs. Keeping two of the three parameters constant we investigate if spiders are small worlds in the sense recently defined by Egghe.

math.CO↗
arXiv2024

Majorization and the degree sequence of trees

We investigate the relation between degree sequences of trees and the majorization order using the Muirhead theorem. In this way, we prove a theorem that provides a necessary and sufficient condition for delta sequences of trees to be comparable in the majorization order. Although our investigation is largely theoretical, our study contributes to a better knowledge of trees as an important data structure. We point out that this study is among the few combining Lorenz curves and majorization on the one hand, and degree sequences of networks on the other.

math.CO↗
arXiv2024

Networks and their degree distribution, leading to a new concept of small worlds

The degree distribution, referred to as the delta-sequence of a network is studied. Using the non-normalized Lorenz curve, we apply a generalized form of the classical majorization partial order. Next, we introduce a new class of small worlds, namely those based on degree centralities of networks. Similar to a previous study, small worlds are defined as sequences of networks with certain limiting properties. We distinguish between three types of small worlds: those based on the highest degree, those based on the average degree, and those based on the median degree. We show that these new classes of small worlds are different from those introduced previously based on the diameter of the network or the average and median distance between nodes. However, there exist sequences of networks that qualify as small worlds in both senses of the word, with stars being an example. Our approach enables the comparison of two networks with an equal number of nodes in terms of their small-worldliness. Finally, we introduced neighboring arrays based on the degrees of the zeroth and first-order neighbors and proved that for trees, equal neighboring arrays lead to equal delta-arrays.

math.GM↗
arXiv2024

The small-world phenomenon: a model, explanations, characterizations and examples

We introduce and define three types of small worlds: small worlds based on the diameter of the network (SWD), those based on the average geodesic distance between nodes (SWA), and those based on the median geodesic distance (SWMd). These types of networks are defined as limiting properties of sequences of sets. We show the exact relation between these three types, namely that each SWD network is also an SWA network and that each SWA network is also an SWMd network. Yet, having the small-world property is rather evident, in the sense that most networks are small-world networks in one of the three ways. We introduce sequences of distance frequencies, so-called alpha-sequences, and prove a relation between the majorization property between alpha-sequences and small-world properties.

cs.SI↗
arXiv2023

Extended Lorenz majorization and frequencies of distances in an undirected network

Findings: We show that the distance distribution in an undirected network Lorenz majorizes the one of a chain. As a consequence, the average and median distances in any such network are smaller than or equal to those of a chain. Research limitations: We restricted our investigations to undirected, unweighted networks. Practical implications: We are convinced that these results are useful in the study of small worlds and the so-called six degrees of separation property.

math.GM↗
arXiv2023

Hirsch-type equations and bundles

We define Hirsch-type equations and bundles being common generalizations of the defining equations of e.g. Hirsch-bundles, g-bundles and Kosmulski-bundles. In this way, common properties of alle these bundles can be proved. The main result proves basic inequalities for these bundles. They form the basis for convergence results as well as for criteria for these bundles to be impact bundles.

math.GM↗
arXiv2023

Hirsch meets Fibonacci and Narayana type variants

For functions $f$ of a continuous variable in $\mathbb{R}^{+}$ we show that the Hirsch function $h_f$ equals $f$ iff $(f(f(x)) = x f(x))$ on $\mathbb{R}^{+}$, leading for continuous $f$ to $f$ = $\emptyset$ or the power function $f(x)$ = $x^α$, $α= \sqrt{5} +1)/2$. For functions of a discrete positive variable in $\mathbb{R}^{+}$, we show that $h_f$ = $f$ implies that only the trivial function $f$ = {(1,1)} satisfies this. We also study the problem $h_f = f \circ f$ and for $f = g \circ g, h_f = g$ leading to the zero function or another power law in the continuous variable case and again to $f$ = {(1,1)} in the discrete variable case. Both problems involve the study of variants of the Fibonacci sequence for which non-trivial identities are proved and applied in the solution of the above problems.

math.GM↗
arXiv2023

Convergence of impact measures and impact bundles

Purpose: A new point of view in the study of impact is introduced. Approach: Using fundamental theorems in real analysis we study the convergence of well-known impact measures. Findings: We show that pointwise convergence is maintained by all well-known impact bundles (such as the h-, g-, and R-bundle) and that the $μ$-bundle even maintains uniform convergence. Based on these results, a classification of impact bundles is given. Research limitations: As for all impact studies, it is just impossible to study all measures in depth. Practical implications: It is proposed to include convergence properties in the study of impact measures. Originality: This article is the first to present a bundle classification based on convergence properties of impact bundles.

math.GM↗
arXiv2022

The Hirsch function and its properties

The Hirsch function of a given continuous function is a new function depending on a parameter. It exists provided some assumptions are satisfied. If this parameter takes the value one, we obtain the well-known h-index. We prove some properties of the Hirsch function and characterize the shape of general functions that are Hirsch functions. We, moreover, present a formula that enables the calculation of f, given its Hirsch function $h_f$.

math.GM↗
arXiv2022

The generalized e-bundle

In previous work, we introduced the notion of an impact bundle, showing how e.g., the h-index and the g-index can lead to such a bundle. Here we extend the set of impact bundles by a new impact bundle, based on the Zhang e-index. It is, moreover, shown that some other plausible definitions do not lead to an impact bundle.

cs.SI↗
arXiv2022

Impact in informetrics and beyond

The concept of impact is one of the most important concepts in informetrics. It is here studied mathematically. We first fix a topic for which we want to find influential objects such as authors or journals, and their production, such as publications generating citations. These objects are then said to have a certain degree of impact. We work on three levels. On the first level, we need a measure for these objects, represented by their rank-frequency function, describing the number of items per source (ranked in decreasing order of the number of items): an impact measure. These measures focus on the production of the most productive sources. The h-index is one example. In paper II we study a formal definition of impact measures based on these left-hand sides of the source-item rank-frequency functions representing these objects. The second level of impact investigation is using impact bundles (or sheaves) as in paper III. As an illustration, we mention that the h-index of a function Z is defined as x for which Z(x) = x, i.e., the abscissa of the intersection of the graph of Z with the line y = x. The h-bundle is defined in the same way but now the line y = x is replaced by an increasing line through the origin: y = $θ$.x, $θ$ > 0. So, we have a bundle of impact measures which is more powerful to measure the impact of an object Z. Impact bundles are characterized in paper III. A third level of impact investigations involves the non-normalized form of the Lorenz curve. In papers IV and V we study global impact measures as measures that respect the non-normalized Lorenz order between the rank-frequency functions representing objects Z. We say that object Z has more impact than object Y if Y is smaller than Z in the sense of the non-normalized Lorenz order. This is the highest level of impact treatment: a mathematical definition of the concept itself.

cs.DL↗
arXiv2022

Generalized Lorenz dominance orders

We extend the discrete majorization theory by working with non-normalized Lorenz curves. Then we prove two generalizations of the Muirhead theorem. These not only use elementary transfers but also local increases. Together these operations are described as elementary impact increases. The first generalization shows that if an array X is dominated, in the generalized sense, by an array Y then Y can be derived from X by a finite number of elementary impact increases and this in such a way that each step transforms an array into a new one which is strictly larger in the generalized majorization sense. The other one shows that if the dominating array, Y, is ordered decreasingly then elementary impact increases starting from the dominated array, X, lead to the dominating one. Here each step transforms an array to a new one for which the decreasingly ordered version dominates the previous one and is dominated by Y.

cs.DM↗
arXiv2009

The relation between Pearson's correlation coefficient r and Salton's cosine measure

The relation between Pearson's correlation coefficient and Salton's cosine measure is revealed based on the different possible values of the division of the L1-norm and the L2-norm of a vector. These different values yield a sheaf of increasingly straight lines which form together a cloud of points, being the investigated relation. The theoretical results are tested against the author co-citation relations among 24 informetricians for whom two matrices can be constructed, based on co-citations: the asymmetric occurrence matrix and the symmetric co-citation matrix. Both examples completely confirm the theoretical results. The results enable us to specify an algorithm which provides a threshold value for the cosine above which none of the corresponding Pearson correlations would be negative. Using this threshold value can be expected to optimize the visualization of the vector space.

cs.IR↗