KLUEDO RSS FeedNeueste Dokumente / Latest documents
https://kluedo.ub.uni-kl.de/index/index/
Wed, 26 Mar 2014 12:14:40 +0100Wed, 26 Mar 2014 12:14:40 +0100Edgeworth expansions for lattice triangular arrays
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3765
Edgeworth expansions have been introduced as a generalization of the central limit theorem and allow to investigate the convergence properties of sums of i.i.d. random variables. We consider triangular arrays of lattice random vectors and obtain a valid Edgeworth expansion for this case. The presented results can be used, for example, to study the convergence behavior of lattice models.Alona Bockpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3765Wed, 26 Mar 2014 12:14:40 +0100Monitoring time series based on estimating functions
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3693
A large class of estimators including maximum likelihood, least squares and M-estimators are based on estimating functions. In sequential change point detection related monitoring functions can be used to monitor new incoming observations based on an initial estimator, which is computationally efficient because possible numeric optimization is restricted to the initial estimation. In this work, we give general regularity conditions under which we derive the asymptotic null behavior of the corresponding tests in addition to their behavior under alternatives, where conditions become particularly simple for sufficiently smooth estimating and monitoring functions. These regularity conditions unify and even extend a large amount of existing procedures in the literature, while they also allow us to derive monitoring schemes in time series that have not yet been considered in the literature including non-linear autoregressive time series and certain count time series such as binary or Poisson autoregressive models. We do not assume that the estimating and monitoring function are equal or even of the same dimension, allowing for example to combine a non-robust but more precise initial estimator with a robust monitoring scheme. Some simulations and data examples illustrate the usefulness of the described procedures.Claudia Kirch; Joseph Tadjuidje Kamgaingpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3693Wed, 29 Jan 2014 10:20:27 +0100On the Generality of the Greedy Algorithm for Solving Matroid Base Problems
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3535
It is well known that the greedy algorithm solves matroid base problems for all linear cost functions and is, in fact, correct if and only if the underlying combinatorial structure of the problem is a matroid. Moreover, the algorithm can be applied to problems with sum, bottleneck, algebraic sum or \(k\)-sum objective functions. Lara Turner; Matthias Ehrgott; Horst W. Hamacherpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3535Wed, 19 Jun 2013 08:27:31 +0200Maximum Likelihood Estimators for Multivariate Hidden Markov Mixture Models
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3480
In this paper we consider a multivariate switching model, with constant states means
and covariances. In this model, the switching mechanism between the basic states of
the observed time series is controlled by a hidden Markov chain. As illustration, under
Gaussian assumption on the innovations and some rather simple conditions, we prove
the consistency and asymptotic normality of the maximum likelihood estimates of the model parameters.Joseph Tadjuidje Kamgaingpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3480Mon, 15 Apr 2013 17:29:52 +0200A limitation of the estimation of intrinsic volumes via pixel configuration counts
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3273
It is often helpful to compute the intrinsic volumes of a set of which only a pixel image is observed. A computational efficient approach, which is suggested by several authors and used in practice, is to approximate the intrinsic volumes by a linear functional of the pixel configuration histogram. Here we want to examine, whether there is an optimal way of choosing this linear functional, where we will use a quite natural optimality criterion that has already been applied successfully for the estimation of the surface area. We will see that for intrinsic volumes other than volume or surface area this optimality criterion cannot be used, since estimators which ignore the data and return constant values are optimal w.r.t. this criterion. This shows that one has to be very careful, when intrinsic volumes are approximated by a linear functional of the pixel configuration histogram.Jürgen Kampfpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/3273Mon, 01 Oct 2012 15:52:05 +0200Asymptotic Order of the Parallel Volume Difference
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2912
In this paper we investigate the asymptotic behaviour of the parallel volume
of fixed non-convex bodies in Minkowski spaces as the distance \(r\) tends to infinity.
We will show that the difference of the parallel volume of the convex hull of a
body and the parallel volume of the body itself can at most have order \(r^{d-2}\) in a \(d\)-dimensional space. Then we will show that in Euclidean spaces this difference can at most have order \(r^{d-3}\). These results have several applications, e.g. we will use
them to compute the derivative of \(f_\mu(rK)\) in \(r = 0\), where \(f_\mu\) is the Wills functional
or a similar functional, \(K\) is a body and \(rK\) is the Minkowski-product of \(r\) and \(K\). Finally we present applications concerning Brownian paths and Boolean models and derive new proofs for formulae for intrinsic volumes.Jürgen Kampfpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2912Mon, 27 Feb 2012 10:09:25 +0100Changepoint tests for INARCH time series
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2725
In this paper, we discuss the problem of testing for a changepoint in the structure
of an integer-valued time series. In particular, we consider a test statistic
of cumulative sum (CUSUM) type for general Poisson autoregressions of order
1. We investigate the asymptotic behaviour of conditional least-squares estimates
of the parameters in the presence of a changepoint. Then, we derive the
asymptotic distribution of the test statistic under the hypothesis of no change,
allowing for the calculation of critical values. We prove consistency of the test,
i.e. asymptotic power 1, and consistency of the corresponding changepoint estimate.
As an application, we have a look at changepoint detection in daily
epileptic seizure counts from a clinical study.Jürgen Franke; Claudia Kirch; Joseph Tadjuidje Kamgaingpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2725Mon, 12 Sep 2011 09:49:44 +0200Variants of the Shortest Path Problem
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2713
The shortest path problem in which the \((s,t)\)-paths \(P\) of a given digraph \(G =(V,E)\) are compared with respect to the sum of their edge costs is one of the best known problems in combinatorial optimization. The paper is concerned with a number of variations of this problem having different objective functions like bottleneck, balanced, minimum deviation, algebraic sum, \(k\)-sum and \(k\)-max objectives, \((k_1, k_2)-max, (k_1, k_2)\)-balanced and several types of trimmed-mean objectives. We give a survey on existing algorithms and propose a general model for those problems not yet treated in literature. The latter is based on the solution of resource constrained shortest path problems with equality constraints which can be solved in pseudo-polynomial time if the given graph is acyclic and the number of resources is fixed. In our setting, however, these problems can be solved in strongly polynomial time. Combining this with known results on \(k\)-sum and \(k\)-max optimization for general combinatorial problems, we obtain strongly polynomial algorithms for a variety of path problems on acyclic and general digraphs. Lara Turnerpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2713Thu, 25 Aug 2011 09:14:11 +0200Asymptotic order of the parallel volume difference
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2327
In this paper we continue the investigation of the asymptotic behavior of the parallel volume in Minkowski spaces as the distance tends to infinity that was started in [13]. We will show that the difference of the parallel volume of the convex hull of a body and the parallel volume of the body itself can at most have order \(r^{d-2}\) in dimension \(d\). Then we will show that in the Euclidean case this difference can at most have order \(r^{d-3}\). We will also examine the asymptotic behavior of the derivative of this difference as the distance tends to infinity. After this we will compute the derivative of \(f_\mu (rK)\) in \(r\), where \(f_\mu\) is the Wills functional or a similar functional, \(K\) is a fixed body and \(rK\) is the Minkowski-product of \(r\) and \(K\). Finally we will use these results to examine Brownian paths and Boolean models and derive new proofs for formulae for intrinsic volumes.Jürgen Kampfpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2327Thu, 26 May 2011 11:11:40 +0200A uniform central limit theorem for neural network based autoregressive processes with applications to change-point analysis
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2302
We consider an autoregressive process with a nonlinear regression function that is modeled by a feedforward neural network. We derive a uniform central limit theorem which is useful in the context of change-point analysis. We propose a test for a change in the autoregression function which - by the uniform central limit theorem - has asymptotic power one for a large class of alternatives including local alternatives.Claudia Kirch; Joseph Tadjuidje Kamgaingpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2302Fri, 25 Mar 2011 14:44:40 +0100Testing for parameter stability in nonlinear autoregressive models
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2301
In this paper we develop testing procedures for the detection of structural changes in nonlinear autoregressive processes. For the detection procedure we model the regression function by a single layer feedforward neural network. We show that CUSUM-type tests based on cumulative sums of estimated residuals, that have been intensively studied for linear regression, can be extended to this case. The limit distribution under the null hypothesis is obtained, which is needed to construct asymptotic tests. For a large class of alternatives it is shown that the tests have asymptotic power one. In this case, we obtain a consistent change-point estimator which is related to the test statistics. Power and size are further investigated in a small simulation study with a particular emphasis on situations where the model is misspecified, i.e. the data is not generated by a neural network but some other regression function. As illustration, an application on the Nile data set as well as S&P log-returns is given.Claudia Kirch; Joseph Tadjuidje Kamgaingpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2301Fri, 25 Mar 2011 14:44:06 +0100On a Cardinality Constrained Multicriteria Knapsack Problem
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2281
We consider a variant of a knapsack problem with a fixed cardinality constraint. There are three objective functions to be optimized: one real-valued and two integer-valued objectives. We show that this problem can be solved efficiently by a local search. The algorithm utilizes connectedness of a subset of feasible solutions and has optimal run-time.Florian Seipp; Stefan Ruzika; Luis Paquetepreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2281Tue, 08 Feb 2011 12:18:44 +0100Generalized Multiple Objective Bottleneck Problems
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2252
We consider multiple objective combinatiorial optimization problems in which the first objective is of arbitrary type and the remaining objectives are either bottleneck or k-max objective functions. While the objective value of a bottleneck objective is determined by the largest cost value of any element in a feasible solution, the kth-largest element defines the objective value of the k-max objective. An efficient solution approach for the generation of the complete nondominated set is developed which is independent of the specific combinatiorial problem at hand. This implies a polynomial time algorithm for several important problem classes like shortest paths, spanning tree, and assignment problems with bottleneck objectives which are known to be NP-hard in the general multiple objective case.Jochen Gorski; Kathrin Klamroth; Stefan Ruzikapreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2252Thu, 09 Dec 2010 10:21:17 +0100Universal Shortest Paths
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2230
We introduce the universal shortest path problem (Univ-SPP) which generalizes both - classical and new - shortest path problems. Starting with the definition of the even more general universal combinatorial optimization problem (Univ-COP), we show that a variety of objective functions for general combinatorial problems can be modeled if all feasible solutions have the same cardinality. Since this assumption is, in general, not satisfied when considering shortest paths, we give two alternative definitions for Univ-SPP, one based on a sequence of cardinality contrained subproblems, the other using an auxiliary construction to establish uniform length for all paths between source and sink. Both alternatives are shown to be (strongly) NP-hard and they can be formulated as quadratic integer or mixed integer linear programs. On graphs with specific assumptions on edge costs and path lengths, the second version of Univ-SPP can be solved as classical sum shortest path problem.Lara Turner; Horst W. Hamacherpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2230Mon, 16 Aug 2010 11:00:30 +0200Weak Dependence of Functional INGARCH Processes
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2186
We introduce a class of models for time series of counts which include INGARCH-type models as well as log linear models for conditionally Poisson distributed data. For those processes, we formulate simple conditions for stationarity and weak dependence with a geometric rate. The coupling argument used in the proof serves as a role model for a similar treatment of integer-valued time series models based on other types of thinning operations.Jürgen Frankepreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2186Thu, 15 Apr 2010 07:56:55 +0200Maximum Likelihood Estimators for Markov Switching Autoregressive Processes with ARCH Component
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2146
We consider a mixture of AR-ARCH models where the switching between the basic states of the observed time series is controlled by a hidden Markov chain. Under simple conditions, we prove consistency and asymptotic normality of the maximum likelihood parameter estimates combining general results on asymptotics of Douc et al (2004) and of geometric ergodicity of Franke et al (2007).Jürgen Franke; Joseph Tadjuidje Kamgaingpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2146Mon, 19 Oct 2009 17:01:13 +0200Mixtures of Nonparametric Autoregression, revised
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2115
We consider data generating mechanisms which can be represented as mixtures of finitely many regression or autoregression models. We propose nonparametric estimators for the functions characterizing the various mixture components based on a local quasi maximum likelihood approach and prove their consistency. We present an EM algorithm for calculating the estimates numerically which is mainly based on iteratively applying common local smoothers and discuss its convergence properties.Jürgen Franke; Jean-Pierre Stockis; Joseph Tadjuidje; W.K. Lipreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2115Mon, 27 Jul 2009 08:47:55 +0200Mixtures of Nonparametric Autoregressions
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2102
We consider data generating mechanisms which can be represented as mixtures of finitely many regression or autoregression models. We propose nonparametric estimators for the functions characterizing the various mixture components based on a local quasi maximum likelihood approach and prove their consistency. We present an EM algorithm for calculating the estimates numerically which is mainly based on iteratively applying common local smoothers and discuss its convergence properties.Jürgen Franke; Jean-Pierre Stockis; Joseph Tadjuidje; W.K. Lipreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2102Mon, 13 Jul 2009 15:52:26 +0200A Note On Inverse Max Flow Problem Under Chebyshev Norm
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2056
In this paper, we study the inverse maximum flow problem under \(\ell_\infty\)-norm and show that this problem can be solved by finding a maximum capacity path on a modified graph. Moreover, we consider an extension of the problem where we minimize the number of perturbations among all the optimal solutions of Chebyshev norm. This bicriteria version of the inverse maximum flow problem can also be solved in strongly polynomial time by finding a minimum \(s - t\) cut on the modified graph with a new capacity function.Cigdem Güler; Horst W. Hamacherpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2056Thu, 15 Jan 2009 15:19:44 +0100A Class of Switching Regimes Autoregressive Driven Processes with Exogenous Components
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2016
In this paper we develop a data-driven mixture of vector autoregressive models with exogenous components. The process is assumed to change regimes according to an underlying Markov process. In contrast to the hidden Markov setup, we allow the transition probabilities of the underlying Markov process to depend on past time series values and exogenous variables. Such processes have potential applications to modeling brain signals. For example, brain activity at time t (measured by electroencephalograms) will can be modeled as a function of both its past values as well as exogenous variables (such as visual or somatosensory stimuli). Furthermore, we establish stationarity, geometric ergodicity and the existence of moments for these processes under suitable conditions on the parameters of the model. Such properties are important for understanding the stability properties of the model as well as deriving the asymptotic behavior of various statistics and model parameter estimators.Joseph Tadjuidje Kamgaing; Hernando Ombao; Richard A. Davispreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2016Wed, 23 Jul 2008 15:16:26 +0200Inverse Tension Problems and Monotropic Optimization
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2015
Given a directed graph G = (N,A), a tension is a function from A to R which satisfies Kirchhoff\\\'s law for voltages. There are two well-known tension problems on graphs. In the minimum cost tension problem (MCT), a cost vector is given and a tension satisfying lower and upper bounds is seeked such that the total cost is minimum. In the maximum tension problem (MaxT), the graph contains 2 special nodes and an arc between them. The aim is to find the maximum tension on this arc. In this study we assume that both problems are feasible and have finite optimal solutions and analyze their inverse versions under rectilinear and Chebyshev distances. In the inverse minimum cost tension problem we adjust the cost parameter to make a given feasible solution the optimum, whereas in inverse maximum tension problem the bounds of the arcs are modified. We show, by extending the results of Ahuja and Orlin (2002), that these inverse tension problems are in a way \\\"dual\\\" to the inverse network flows. We prove that the inverse minimum cost tension problem under rectilinear norm is equivalent to solving a minimum cost tension problem, while under unit weight Chebyshev norm it can be solved by finding a minimum mean cost residual cut. Moreover, inverse maximum tension problem under rectilinear norm can be solved as a maximum tension problem on the same graph with new arc bounds. Finally, we provide a generalization of the inverse problems to monotropic programming problems with linear costs.Cigdem Gülerpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/2015Wed, 23 Jul 2008 15:16:20 +0200A new sequential extraction heuristic for optimising the delivery of cancer radiation treatment using multileaf collimators
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1930
Finding a delivery plan for cancer radiation treatment using multileaf collimators operating in ''step-and-shoot mode'' can be formulated mathematically as a problem of decomposing an integer matrix into a weighted sum of binary matrices having the consecutive-ones property - and sometimes other properties related to the collimator technology. The efficiency of the delivery plan is measured by both the sum of weights in the decomposition, known as the total beam-on time, and the number of different binary matrices appearing in it, referred to as the cardinality, the latter being closely related to the set-up time of the treatment. In practice, the total beam-on time is usually restricted to its minimum possible value, (which is easy to find), and a decomposition that minimises cardinality (subject to this restriction) is sought.Davaasteren Baatar; Natashia Boland; Robert Johnston; Horst W. Hamacherpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1930Fri, 25 Jan 2008 22:43:20 +0100Capacity Inverse Minimum Cost Flow Problem
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1914
Given a directed graph G = (N,A) with arc capacities u and a minimum cost flow problem defined on G, the capacity inverse minimum cost flow problem is to find a new capacity vector u' for the arc set A such that a given feasible flow x' is optimal with respect to the modified capacities. Among all capacity vectors u' satisfying this condition, we would like to find one with minimum ||u' - u|| value. We consider two distance measures for ||u' - u||, rectilinear and Chebyshev distances. By reduction from the feedback arc set problem we show that the capacity inverse minimum cost flow problem is NP-hard in the rectilinear case. On the other hand, it is polynomially solvable by a greedy algorithm for the Chebyshev norm. In the latter case we propose a heuristic for the bicriteria problem, where we minimize among all optimal solutions the number of affected arcs. We also present computational results for this heuristic.Cigdem Güler; Horst Hamacherpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1914Mon, 03 Dec 2007 12:01:57 +0100Some asymptotics for local least-squares regression with regularization
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1902
We derive some asymptotics for a new approach to curve estimation proposed by Mr'{a}zek et al. cite{MWB06} which combines localization and regularization. This methodology has been considered as the basis of a unified framework covering various different smoothing methods in the analogous two-dimensional problem of image denoising. As a first step for understanding this approach theoretically, we restrict our discussion here to the least-squares distance where we have explicit formulas for the function estimates and where we can derive a rather complete asymptotic theory from known results for the Priestley-Chao curve estimate. In this paper, we consider only the case where the bias dominates the mean-square error. Other situations are dealt with in subsequent papers.Jürgen Franke; Joseph Tadjuidje; Stefan Didas; Joachim Weickertpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1902Thu, 11 Oct 2007 12:37:44 +0200New heuristics for the minimum fundamental cut basis problem
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1887
Given an undirected connected network and a weight function finding a basis of the cut space with minimum sum of the cut weights is termed Minimum Cut Basis Problem. This problem can be solved, e.g., by the algorithm of Gomory and Hu [GH61]. If, however, fundamentality is required, i.e., the basis is induced by a spanning tree T in G, the problem becomes NP-hard. Theoretical and numerical results on that topic can be found in Bunke et al. [BHMM07] and in Bunke [Bun06]. In the following we present heuristics with complexity O(m log n) and O(mn), where n and m are the numbers of vertices and edges respectively, which obtain upper bounds on the aforementioned problem and in several cases outperform the heuristics of Schwahn [Sch05].Alexander J. Perez Tchernov; Anne M. Schwahnpreprinthttps://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/1887Tue, 24 Jul 2007 14:31:16 +0200