Logistic biplot for nominal data

Classical biplot methods allow for the simultaneous representation of individuals (rows) and variables (columns) of a data matrix. For binary data, logistic biplots have been recently developed. When data are nominal, both classical and binary logistic biplots are not adequate and techniques such as multiple correspondence analysis (MCA), latent trait analysis (LTA) or item response theory (IRT) for nominal items should be used instead. In this paper we extend the binary logistic biplot to nominal data. The resulting method is termed “nominal logistic biplot”(NLB), although the variables are represented as convex prediction regions rather than vectors. Using the methods from computational geometry, the set of prediction regions is converted to a set of points in such a way that the prediction for each individual is established by its closest “category point”. Then interpretation is based on distances rather than on projections. We study the geometry of such a representation and construct computational algorithms for the estimation of parameters and the calculation of prediction regions. Nominal logistic biplots extend both MCA and LTA in the sense that they give a graphical representation for LTA similar to the one obtained in MCA.     

There is no (46, 6, 1) block design*

Abstact: In this paper we show that a (46, 6, 1) design does not exist. This result was obtained by a computer search. In the incidence matrix of such a design, there must exist a “c4” configuration—6 rows and 4 columns, in which each pair of columns intersect exactly once, in distinct rows. There can also exist a “c5” configuration with 10 rows and 5 columns, in which each pair of columns intersect exactly once, in distinct rows. Thus the search for (46, 6, 1) designs can be subdivided into two cases, the first of which assumes there is no “c5”, and the second of which assumes there is a “c5”. After completing the searches for both cases, we found no (46, 6, 1) design. © 2000 John Wiley & Sons, Inc. J Combin Designs 9: 60–71, 2001     

A complete analysis of Black-White Hackendot

RecentlyÚlehla [1980] gave a complete analysis of an impartial two-person combinatorial game called Hackendot which was invented by von Neumann. In this note we consider a partizan (or unimpartial) variation of Hackendot. Úlehla's analysis uses the Sprague-Grundy theory for disjunctive sums of impartial games. Our analysis employs Conway's theory for disjunctive sums of partizan games, two certain “biased” ways of “adding” dyadic rationals, and the usual addition of numbers.     

FACTORIZATIONS OF CONES ALONG A FUNCTOR

Abstract

First a general Galois correspondence is established, which generalizes at the same time the correspondence between classes of monomorphisms and injective objects and the correspondence between classes of epimorphisms and monomorphisms in a category. This correspondence arises naturally if one tries to generalize some concepts of “topological” or also of “algebraic” functors. Both kinds of functors admit certain factorizations of cones, and just this fact implies some of their common nice properties: lifting limits, continuity and faithfulness, for instance. These properties can be shown without having a left adjoint. Therefore the theory yields also applications to functors which are neither “topological” nor “algebraic”.     

Clustering of contingency table and mixture model

Basing cluster analysis on mixture models has become a classical and powerful approach. It enables some classical criteria such as the well-known k-means criterion to be explained. To classify the rows or the columns of a contingency table, an adapted version of k-means known as Mndki2, which uses the chi-square distance, can be used. Unfortunately, this simple, effective method which can be used jointly with correspondence analysis based on the same representation of the data, cannot be associated with a mixture model in the same way as the classical k-means algorithm. In this paper we show that the Mndki2 algorithm can be viewed as an approximation of a classifying version of the EM algorithm for a mixture of multinomial distributions. A comparison of the algorithms belonging in this context are experimentally investigated using Monte Carlo simulations.     

Discovery of time-inconsecutive co-movement patterns of foreign currencies using an evolutionary biclustering method

This paper proposes an evolutionary biclustering algorithm to discover inconsecutive co-movement patterns of different foreign exchange rates. The rows/columns of a bicluster (i.e. a submatrix with a subset of rows and a subset of columns) are not necessarily consecutive. A typical bicluster with constant values on rows and/or columns is represented as a hyperplane in a high-dimensional space and the coefficients of the hyperplane are determined using a genetic algorithm. A detected bicluster demonstrates the co-moving behaviors of a subset of currencies in inconsecutive time periods, indicating that the currencies moved in different manners in some specific time periods. In our experiments, we relate these patterns to the geographically close economic connections and find out the correspondence between the nominal exchange rates and the economic conditions. The findings are useful as a guide for investing foreign currencies.     

The phase transition in inhomogeneous random graphs

The “classical” random graph models, in particular G(n,p), are “homogeneous,” in the sense that the degrees (for example) tend to be concentrated around a typical value. Many graphs arising in the real world do not have this property, having, for example, power‐law degree distributions. Thus there has been a lot of recent interest in defining and studying “inhomogeneous” random graph models. One of the most studied properties of these new models is their “robustness”, or, equivalently, the “phase transition” as an edge density parameter is varied. For G(n,p), p = c/n, the phase transition at c = 1 has been a central topic in the study of random graphs for well over 40 years. Many of the new inhomogeneous models are rather complicated; although there are exceptions, in most cases precise questions such as determining exactly the critical point of the phase transition are approachable only when there is independence between the edges. Fortunately, some models studied have this property already, and others can be approximated by models with independence. Here we introduce a very general model of an inhomogeneous random graph with (conditional) independence between the edges, which scales so that the number of edges is linear in the number of vertices. This scaling corresponds to the p = c/n scaling for G(n,p) used to study the phase transition; also, it seems to be a property of many large real‐world graphs. Our model includes as special cases many models previously studied. We show that, under one very weak assumption (that the expected number of edges is “what it should be”), many properties of the model can be determined, in particular the critical point of the phase transition, and the size of the giant component above the transition. We do this by relating our random graphs to branching processes, which are much easier to analyze. We also consider other properties of the model, showing, for example, that when there is a giant component, it is “stable”: for a typical random graph, no matter how we add or delete o(n) edges, the size of the giant component does not change by more than o(n). © 2007 Wiley Periodicals, Inc. Random Struct. Alg., 31, 3–122, 2007     

Pointwise convergence to the initial data for nonlocal dyadic diffusions

We solve the initial value problem for the diffusion induced by dyadic fractional derivative s in ?⁺. First we obtain the spectral analysis of the dyadic fractional derivative operator in terms of the Haar system, which unveils a structure for the underlying “heat kernel”. We show that this kernel admits an integrable and decreasing majorant that involves the dyadic distance. This allows us to provide an estimate of the maximal operator of the diffusion by the Hardy-Littlewood dyadic maximal operator. As a consequence we obtain the pointwise convergence to the initial data.     

The bijection between plane partitions and nonnegative integer matrices

The one-to-one correspondence between the set of plane partitions withr rows andm columns and the set of matrices of nonnegative integers with the same numbers of rows and columns has been constructed. Published in Lietuvos Matematikes Rinkinys, Vol. 35, No. 2, pp. 204–210, April–June, 1995.     

Inferring social structure from continuous‐time interaction data

Relational event data, which consist of events involving pairs of actors over time, are now commonly available at the finest of temporal resolutions. Existing continuous‐time methods for modeling such data are based on point processes and directly model interaction “contagion,” whereby one interaction increases the propensity of future interactions among actors, often as dictated by some latent variable structure. In this article, we present an alternative approach to using temporal‐relational point process models for continuous‐time event data. We characterize interactions between a pair of actors as either spurious or as resulting from an underlying, persistent connection in a latent social network. We argue that consistent deviations from expected behavior, rather than solely high frequency counts, are crucial for identifying well‐established underlying social relationships. This study aims to explore these latent network structures in two contexts: one comprising of college students and another involving barn swallows.     

A Characterization of Ordered Sets and Lattices via Betweenness Relations

For a set X with at least 3 elements, we establish a canonical one to one correspondence between all betweenness relations satisfying certain axioms and all pairs of inverse orderings “<” and “>” defined on X for which the corresponding Hasse diagram is connected and all maximal chains contain at least 3 elements. For an ordering “<”, the corresponding betweenness relation B is given by $$B=\{(x,y,z)\in X^3\mid x<y<z {\rm \ or\ }z<y<x\}.$$ Moreover, by adding one more axiom, we obtain also a one to one correspondence between all pairs of dual lattices and all betweenness relations.     

On the complexity of the dualization problem

The computational complexity of discrete problems concerning the enumeration of solutions is addressed. The concept of an asymptotically efficient algorithm is introduced for the dualization problem, which is formulated as the problem of constructing irreducible coverings of a Boolean matrix. This concept imposes weaker constraints on the number of “redundant” algorithmic steps as compared with the previously introduced concept of an asymptotically optimal algorithm. When the number of rows in a Boolean matrix is no less than the number of columns (in which case asymptotically optimal algorithms for the problem fail to be constructed), algorithms based on the polynomialtime-delay enumeration of “compatible” sets of columns of the matrix is shown to be asymptotically efficient. A similar result is obtained for the problem of searching for maximal conjunctions of a monotone Boolean function defined by a conjunctive normal form.     

Monads with arities and their associated theories

After a review of the concept of “monad with arities” we show that the category of algebras for such a monad has a canonical dense generator. This is used to extend the correspondence between finitary monads on sets and Lawvere’s algebraic theories to a general correspondence between monads and theories for a given category with arities. As an application we determine arities for the free groupoid monad on involutive graphs and recover the symmetric simplicial nerve characterisation of groupoids.     

A Monte Carlo evaluation of three methods to detect local dependence in binary data latent class models

Binary data latent class analysis is a form of model-based clustering applied in a wide range of fields. A central assumption of this model is that of conditional independence of responses given latent class membership, often referred to as the “local independence” assumption. The results of latent class analysis may be severely biased when this crucial assumption is violated; investigating the degree to which bivariate relationships between observed variables fit this hypothesis therefore provides vital information. This article evaluates three methods of doing so. The first is the commonly applied method of referring the so-called “bivariate residuals” to a Chi-square distribution. We also introduce two alternative methods that are novel to the investigation of local dependence in latent class analysis: bootstrapping the bivariate residuals, and the asymptotic score test or “modification index”. Our Monte Carlo simulation indicates that the latter two methods perform adequately, while the first method does not perform as intended.     

Dirac Mass Dynamics in Multidimensional Nonlocal Parabolic Equations

Nonlocal Lotka–Volterra models have the property that solutions concentrate as Dirac masses in the limit of small diffusion. Is it possible to describe the dynamics of the limiting concentration points and of the weights of the Dirac masses? What is the long time asymptotics of these Dirac masses? Can several Dirac masses co-exist?

We will explain how these questions relate to the so-called “constrained Hamilton–Jacobi equation” and how a form of canonical equation can be established. This equation has been established assuming smoothness. Here we build a framework where smooth solutions exist and thus the full theory can be developed rigorously. We also show that our form of canonical equation comes with a kind of Lyapunov functional.

Numerical simulations show that the trajectories can exhibit unexpected dynamics well explained by this equation.

Our motivation comes from population adaptive evolution a branch of mathematical ecology which models Darwinian evolution.     

Numerical Bifurcation and Spectral Stability of Wavetrains in Bidirectional Whitham Models

We consider several different bidirectional Whitham equations that have recently appeared in the literature. Each of these models combines the full two‐way dispersion relation from the incompressible Euler equations with a canonical shallow water nonlinearity, providing nonlocal model equations that may be expected to exhibit some of the interesting high‐frequency phenomena present in the Euler equations that standard “long‐wave” theories fail to capture. Of particular interest here is the existence and stability of periodic traveling wave solutions in such models. Using numerical bifurcation techniques, we construct global bifurcation diagrams for each system and compare the global structure of branches, together with the possibility of bifurcation branches terminating in a “highest” singular (peaked/cusped) wave. We also numerically approximate the stability spectrum along these bifurcation branches and compare the stability predictions of these models. Our results confirm a number of analytical results concerning the stability of asymptotically small waves in these models and provide new insights into the existence and stability of large amplitude waves.     

RUMINATIONS ON THE DEVELOPMENT AND FUTURE OF POPULATION DYNAMICS MODELS IN FISHERIES

ABSTRACT. I trace the development of fisheries models (i.e., fish population dynamics models of species subject to fisheries) to the 21st century. The first real efforts occurred in the period 1900 1920 with the work of Baranov (the “Grandfather” of fisheries population dynamics) and the formation of the International Council for the Exploration of the Sea (ICES). The establishment of the science occurred between 1920 1960 with multi‐species modeling, age‐ and size‐structure dynamics, and production models. Fundamental work during this time was done by Ricker (the “Father” of fisheries population dynamics), Beverton and Holt (the “Prophets” of fisheries population dynamics), Chapman, Dickie, DeLury, Graham, Gulland, Leslie, Lotka and Volterra, Russell, Schaefer, and Thompson. During this time, most of the workwas deterministic and mathematical. Between 1960 and 1980, statistical methodology evolved greatly but was separate from mathematical advances for the most part. The development of statistical principles for the estimation of animal abundance was further enhanced by Arnason, Buckland, Burnham and Anderson and White, Cormack, Eberhardt, Jolly, Manly, Pollock, Ricker, Robson, and Seber, among others. Fisheries models evolved in a deterministic setting, with advances in age‐structured models (Gulland, Pope, Doubleday), surplus production models (Pella, Tomlin‐son, Schnute, Fletcher, Hilborn), growth models, bioeconomic models (C. Clark) and management control models (Hilborn, Walters). The period 1980 2000 was the Golden Age. The integration between mathematics and statistics occurred when likelihood and least squares techniques were formally combined with mathematical models of population change. The number of fisheries modelers grew exponentially during this time, resulting in a concomitant increase in publications. A major advance in the 1990s has been the development of Bayesian and time series methods, which have allowed explicit specification of uncertainty. Currently, theory allows realistic modeling of age‐ and size‐structured populations, migratory populations and harvesting strategies. These models routinely incorporate measurement error, process error (stochasticity) and time variation. But data needs often overwhelm the performance of models, and greater demands are being placed on models to answer complex questions. There has been poor communication between fisheries and ecological modelers, between fisheries researchers and statisticians, and among fisheries researchers in different geographic locales. Future models will need to deal better with habitat and spatial concerns, genetics, multispecies interactions, environmental factors, effects of harvesting on the ecosystem, model misspecification and so‐cioeconomic concerns. Meta‐analysis, retrospective analysis and operating models are some modern approaches for dealing with uncertainty and providing for sustainable fisheries. However, I fear that current attacks on single‐species models and management may result in rejection of these advances and an attempt to substitute a less scientific approach.     

Data Visualization With Multidimensional Scaling

We discuss methodology for multidimensional scaling (MDS) and its implementation in two software systems, GGvis and XGvis. MDS is a visualization technique for proximity data, that is, data in the form of N × N dissimilarity matrices. MDS constructs maps (“configurations,” “embeddings”) in IR^k by interpreting the dissimilarities as distances. Two frequent sources of dissimilarities are high-dimensional data and graphs. When the dissimilarities are distances between high-dimensional objects, MDS acts as a (often nonlinear) dimension-reduction technique. When the dissimilarities are shortest-path distances in a graph, MDS acts as a graph layout technique. MDS has found recent attention in machine learning motivated by image databases (“Isomap”). MDS is also of interest in view of the popularity of “kernelizing” approaches inspired by Support Vector Machines (SVMs; “kernel PCA”).

This article discusses the following general topics: (1) the stability and multiplicity of MDS solutions; (2) the analysis of structure within and between subsets of objects with missing value schemes in dissimilarity matrices; (3) gradient descent for optimizing general MDS loss functions (“Strain” and “Stress”); (4) a unification of classical (Strain-based) and distance (Stress-based) MDS.

Particular topics include the following: (1) blending of automatic optimization with interactive displacement of configuration points to assist in the search for global optima; (2) forming groups of objects with interactive brushing to create patterned missing values in MDS loss functions; (3) optimizing MDS loss functions for large numbers of objects relative to a small set of anchor points (“external unfolding”); and (4) a non-metric version of classical MDS.

We show applications to the mapping of computer usage data, to the dimension reduction of marketing segmentation data, to the layout of mathematical graphs and social networks, and finally to the spatial reconstruction of molecules.     

Ancestors and descendants in evolving k‐tree models

We consider several random graph models based on k‐trees, which can be generated by applying the probabilistic growth rules “uniform attachment”, “preferential attachment”, or a “saturation”‐rule, respectively, but which also can be described in a combinatorial way. For all of these models we study the number of ancestors and the number of descendants of nodes in the graph by carrying out a precise analysis which leads to exact and limiting distributional results. © 2014 Wiley Periodicals, Inc. Random Struct. Alg. 44, 465–489, 2014