BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Fächergruppe Mathematik und Statistik - ECPv6.17.0//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-ORIGINAL-URL:https://www.hsu-hh.de/statistik
X-WR-CALDESC:Veranstaltungen für Fächergruppe Mathematik und Statistik
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:Europe/Berlin
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20170326T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20171029T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20180325T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20181028T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20190331T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20191027T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20200329T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20201025T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20210328T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20211031T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20220327T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20221030T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20230326T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20231029T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20240331T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20241027T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20250330T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20251026T010000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20240116T154500
DTEND;TZID=Europe/Berlin:20240116T171500
DTSTAMP:20231206T161610Z
CREATED:20231206T161437Z
LAST-MODIFIED:20231206T161610Z
UID:1117-1705419900-1705425300@www.hsu-hh.de
SUMMARY:Andreas Galka (WTD 71 Kiel)
DESCRIPTION:Analysis of Hydroacoustical Time Series by State Space Modelling\nThis talk deals with predictive time-domain modelling of time series\, either univariate or multivariate. The aim is to identify and reconstruct independent source components from the time series. For this purpose\, a special class of linear state space models is employed\, which describes the individual source components by autoregressive moving-average (ARMA) processes. As a useful result of ARMA modelling\, parametric estimates of the power spectra of the source components can be obtained. In a simulation study\, it is demonstrated that also components generated by nonlinear processes can be separated by linear state space modelling\, using Kalman filtering and smoothing. Furthermore\, comparison with algorithms from Independent Component Analysis (ICA) shows that ICA fails to distinguish correlations due to mixing from correlations due to finite data set size\, leading to distorted estimates. Parameter estimation is performed by numerical maximisation of the logarithmic likelihood\, using the Expectation Maximisation (EM) algorithm and other algorithms for optimisation. The implementation of the EM algorithm for state space modelling under constraints is discussed in detail\, and some new results are presented. Finally\, results for applying state space modelling to several hydroacoustic time series are presented\, including components of the underwater signature of ships and vocalisations of marine mammals. Special emphasis is given to phenomena typical of ship components\, such as frequency combs and Doppler effects.
URL:https://www.hsu-hh.de/statistik/event/andreas-galka-wtd-71-kiel
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20240109T154500
DTEND;TZID=Europe/Berlin:20240109T171500
DTSTAMP:20231219T131625Z
CREATED:20231219T131546Z
LAST-MODIFIED:20231219T131625Z
UID:1125-1704815100-1704820500@www.hsu-hh.de
SUMMARY:Christian Weiß (HSU)
DESCRIPTION:Testing for Dependence by Using Ordinal Patterns: an Introduction\nAbout 20 years ago\, ordinal patterns have been introduced as a simple\, robust\, and flexible tool for analyzing the serial dependence structure of univariate real-valued stochastic processes. If applied to continuously distributed processes\, one can derive non-parametric tests of the null hypothesis that the process is independent and identically distributed. Recently\, also more sophisticated tasks for dependence tests have been considered: serial dependence in a discrete-valued process\, the sequential monitoring of serial dependence\, cross-dependence in a multivariate process\, and spatial dependence in a random field. This talk provides an introduction to the aforementioned topics together with illustrative examples\, and it concludes by outlining perspectives for future research.
URL:https://www.hsu-hh.de/statistik/event/christian-weiss-hsu-2024
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20231219T154500
DTEND;TZID=Europe/Berlin:20231219T171500
DTSTAMP:20231206T161553Z
CREATED:20231206T155926Z
LAST-MODIFIED:20231206T161553Z
UID:1108-1703000700-1703006100@www.hsu-hh.de
SUMMARY:Russell Shinohara (University of Pennsylvania)
DESCRIPTION:Statistical Approaches to Harmonization in Multi-Center Medical Imaging Studies\nWhile magnetic resonance imaging (MRI) studies are critical for the diagnosis\, monitoring\, and study of a wide variety of diseases\, their use in quantitative analysis can be complex. An increasingly recognized issue involves the differences between MRI scanners that are used in large multi-center studies. To address this\, the current state of the art is to „regress out“ or „adjust for“ scanner differences. The field has found these methods to be insufficient and has advocated for the adaptation of methods pioneered in genomics to help mitigate inter-scanner differences\, which can vary across the brain and result in both mean and variance shifts. We further study the implications of differences in correlation structures across and between images\, and how this affects downstream inference.
URL:https://www.hsu-hh.de/statistik/event/russell-shinohara-university-of-pennsylvania
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20231212T154500
DTEND;TZID=Europe/Berlin:20231212T171500
DTSTAMP:20231123T091604Z
CREATED:20231123T091604Z
LAST-MODIFIED:20231123T091604Z
UID:1100-1702395900-1702401300@www.hsu-hh.de
SUMMARY:Paul Doukhan (Université Cergy Paris)
DESCRIPTION:Dependence\, examples and tools\nThe talk first aims at recalling some basic facts linking independence and orthogonality structures. This will mange to provide simple conditions to model stochastic dependences. Some other simple features will also suggest a definition of weak proposed in a paper of 1999 with Sana Louhichi which has some advantages with respect to the more classical strong mixing ones as Rosenblatt’s 1956 proposed. Definitely I don’t think that only such conditions have to be used\, since developing some examples\, many other notions of dependence should in fact be used rather than only one. The main focus is definitely the class of models of time series of interest. A kind of botanic description will make it clear that from the point of view of examples\, such weak dependence conditions (also developed in a LNS volume (with other coauthors) admits some nice features as well a a reasonable limit theory. Especially the case of integer valued models of time series has some good connection to weak dependence.
URL:https://www.hsu-hh.de/statistik/event/paul-doukhan-universite-cergy-paris
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20231017T154500
DTEND;TZID=Europe/Berlin:20231017T171500
DTSTAMP:20230919T125843Z
CREATED:20230809T154235Z
LAST-MODIFIED:20230919T125843Z
UID:1081-1697557500-1697562900@www.hsu-hh.de
SUMMARY:Maxime Faymonville (TU Dortmund)
DESCRIPTION:Goodness-of-fit testing for INAR models\nIn recent years\, there has been a growing interest in the analysis of time series of counts. Among the various models designed for dependent count data\, integer-valued autoregressive (INAR) processes enjoy great popularity. These processes serve as a natural extension of the widely known AR model used in the context of continuous autoregressive time series and have been used extensively in the statistical literature. Typically\, statistical inference for INAR models relies on asymptotic theory and tends to rest upon rather stringent (parametric) model assumptions. Notably\, the Poisson-INAR(1) model\, a prominent example\, has received considerable attention in existing literature. We present a novel semiparametric goodness-of-fit test tailored for the INAR model class\, without imposing any parametric assumptions on the distribution of innovations. While parametric assumptions streamline the approach and other straightforward testing strategies\, they often introduce too restrictive model assumptions. Our proposed procedure relies on the specific structure of the joint probability generating function of INAR models. This approach allows for enhancing the versatility and applicability of INAR models by accommodating a broader array of innovation distributions. We prove the validity of our testing procedure and carefully examine its performance characteristics\, including power and size\, through diverse simulation scenarios.
URL:https://www.hsu-hh.de/statistik/event/maxime-faymonville-tu-dortmund
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20231010T154500
DTEND;TZID=Europe/Berlin:20231010T171500
DTSTAMP:20230919T125721Z
CREATED:20230809T153614Z
LAST-MODIFIED:20230919T125721Z
UID:1076-1696952700-1696958100@www.hsu-hh.de
SUMMARY:Angelika Silbernagel (Uni Siegen)
DESCRIPTION:Ordinal patterns: Different representations and their application in the context of dependence\nSince the seminal paper by Bandt and Pompe\, so called ordinal patterns have been used extensively in contexts of data analysis\, dynamical systems as well as time series analysis and mathematical statistics. Ordinal patterns are defined as the description of the order of the values in a data set (vector) of length d. Due to this simple definition\, there are various ways on how to encode ordinal patterns. Sometimes authors pick a certain representation\, telling their readers\, why it is useful in the context they have in mind. Most of the time\, however\, one gets the impression\, that the representation was chosen randomly or only because ‘others have used it before’. \nTherefore\, here we describe and analyze different approaches to represent ordinal patterns. All of these can be found in the literature. The most important representations (plus sub-classes) are compared in terms of their applicability in different contexts\, namely\, we consider digital implementation\, inverse patterns and ties between values. Thereafter\, we consider a simulation study with regard to dependence between time series using so called multivariate ordinal patterns in order to demonstrate the use of ordinal pattern representations in practice. At the end we provide a guideline on which occasions which representation should be used.
URL:https://www.hsu-hh.de/statistik/event/angelika-silbernagel-uni-siegen
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20230621T154500
DTEND;TZID=Europe/Berlin:20230621T171500
DTSTAMP:20230522T082303Z
CREATED:20230522T082303Z
LAST-MODIFIED:20230522T082303Z
UID:1059-1687362300-1687367700@www.hsu-hh.de
SUMMARY:Hakam Kondakji (HSU)
DESCRIPTION:Optimale Portfolios in einem Finanzmarkt mit Gaußscher Drift und Expertenmeinungen\nWir untersuchen optimale Portfoliostrategien für nutzenmaximierende Investoren in einem zeitstetigen Finanzmarktmodell\, bei dem die Driftdurch einen Ornstein-Uhlenbeck-Prozess modelliert wird\, welcher nicht direkt beobachtbar und vom Investor aus den ihm zur Verfügung stehenden Information zu schätzen ist. In der Praxis beziehen Investoren für die Bestimmung ihrer Anlagestrategien auch andere externe Informationsquellen ein\, um die Einschätzung der unbekannten Drift zu verbessern. Dabei handelt es sich z.B. um Wirtschaftsnachrichten\, Unternehmensberichte\, Ratings\, Empfehlungen von Finanzanalysten oder die eigene individuelle Sicht auf die zukünftige Preisentwicklung. Solche externen Informationen werden in der Literatur Expertenmeinungen genannt.
URL:https://www.hsu-hh.de/statistik/event/hakam-kondakji-hsu
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20230614T154500
DTEND;TZID=Europe/Berlin:20230614T171500
DTSTAMP:20230512T094705Z
CREATED:20230512T094704Z
LAST-MODIFIED:20230512T094705Z
UID:1051-1686757500-1686762900@www.hsu-hh.de
SUMMARY:Christian Weiß (HSU) 2023
DESCRIPTION:Über „magische Steine“ in der Statistik\nAuf Charles Stein geht die Idee zurück\, parametrische Verteilungsfamilien auf eindeutige Weise durch eine auf Momenten basierende Identität zu charakterisieren\, wobei eine solche Identität von einer unbestimmten Funktion f abhängt\, welche aus einer großen Klasse von Funktionen frei wählbar ist. Ursprünglich wurde dieser Ansatz dabei im Rahmen der Wahrscheinlichkeitstheorie entwickelt\, um den Fehler bei Verteilungsapproximationen abzuschätzen. Vor einigen Jahren\, als es um die Vorbereitung des späteren DFG-Projekts „Modelldiagnostik für Zähldatenzeitreihen“ ging\, siehe  \nModelldiagnostik für Zähldatenzeitreihen \n \nhatte ich die Idee\, dass man auf Basis einer solchen Stein-Identität vielleicht auch einen einfachen aber flexiblen Goodness-of-Fit-Test (GoF-Test) kreieren könnte\, der bei passender Wahl der o.g. Funktion f für verschiedenste Alternativszenarien trennscharf sein könnte. Was ich ursprünglich nur als „nette kleine Idee“ unter vielen anderen Möglichkeiten eingeschätzt hatte\, hat sich mittlerweile als eine wahre Wundertüte („It’s magic!“) mit diversen (insbesondere statistischen) Anwendungsmöglichkeiten herausgestellt\, weshalb der Titel dieses Vortrags etwas flapsig von „magischen Steinen“ spricht. Im (hoffentlich) kurzweiligen Vortrag selbst soll es um einige solcher Möglichkeiten gehen\, nämlich jene\, deren ich mir bis dato bewusst geworden bin. Beginnend mit der noch wahrscheinlichkeitstheoretischen Anwendung der Berechnung auch komplizierter Momente\, sollen anschließend drei statistische Anwendungsgebiete kurz vorgestellt werden: GoF-Tests für parametrische Verteilungsfamilien\, Kontrollkarten auf Basis von Stein-Identitäten\, und eine ebenfalls auf Stein-Identitäten basierende verallgemeinerte Momentenschätzung. Die dabei präsentierten Resultate sind teils publiziert\, teils „work in progress“\, und zumeist mit verschiedensten Coautoren entstanden\, insbesondere mit Boris Aleksandrov und Simon Nik.
URL:https://www.hsu-hh.de/statistik/event/christian-weiss-hsu-2023
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20230524T154500
DTEND;TZID=Europe/Berlin:20230524T171500
DTSTAMP:20230512T095942Z
CREATED:20230512T093801Z
LAST-MODIFIED:20230512T095942Z
UID:1042-1684943100-1684948500@www.hsu-hh.de
SUMMARY:Philipp Otto (Uni Hannover)
DESCRIPTION:Statistical process monitoring of artificial neural networks\nThe rapid advancement of models based on artificial intelligence demands innovative monitoring techniques which can operate in real time with low computational costs. In machine learning\, especially if we consider artificial neural networks (ANN)\, the models are often trained in a supervised manner. Consequently\, the learned relationship between the input and the output must remain valid during the model’s deployment. If this stationarity assumption holds\, we can conclude that the ANN generate accurate predictions. Otherwise\, the retraining or rebuilding of the model is required. We propose considering the latent feature representation of the data (called „embedding“) generated by the ANN to determine the time when the data stream starts being nonstationary. In particular\, we monitor embeddings by applying multivariate control charts based on the data depth calculation and normalized ranks. The performance of the introduced method is evaluated by designing a benchmark study with various ANN architectures and different underlying data formats.
URL:https://www.hsu-hh.de/statistik/event/philipp-otto-uni-hannover
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20230517T154500
DTEND;TZID=Europe/Berlin:20230517T171500
DTSTAMP:20230508T075818Z
CREATED:20230505T165623Z
LAST-MODIFIED:20230508T075818Z
UID:1022-1684338300-1684343700@www.hsu-hh.de
SUMMARY:Malte Jahn (HSU)
DESCRIPTION:Regressing on distributions in panel models: The nonlinear effect of temperature on regional economic growth\nA framework is proposed for the situation where certain explanatory variables are available at a higher temporal resolution than the dependent variable. The main idea is to use the moments of the empirical distribution of these variables to construct regressors with the correct resolution. As the moments are likely to display nonlinear marginal and interaction effects\, an artificial neural network regression function is proposed. The usefulness is demonstrated by analyzing the influence of daily temperatures in 260 European NUTS2 regions on the yearly growth of gross value added in these regions in the time period 2000 to 2021. In the particular example\, the model allows for an appropriate assessment of regional economic impacts resulting from (future) changes in the regional temperature distribution (mean AND variance).
URL:https://www.hsu-hh.de/statistik/event/malte-jahn-hsu
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20230426T154500
DTEND;TZID=Europe/Berlin:20230426T171500
DTSTAMP:20230505T165601Z
CREATED:20230505T165529Z
LAST-MODIFIED:20230505T165601Z
UID:1025-1682523900-1682529300@www.hsu-hh.de
SUMMARY:Carina Beering (HSU)
DESCRIPTION:Under weak moment conditions\, we provide a functional central limit theorem (FCLT) for weighted sums of locally stationary processes for two main frameworks which differ in terms of the boundedness of the used function. Since the FCLT itself insinuates beneficial effects of a bootstrap analogue\, we transfer our previous results to the bootstrap world using a local block bootstrap approach. Afterwards\, we combine real- and bootstrap-world findings to present a testing procedure for independence of locally stationary processes using a weighted distance composed of characteristic functions and its empirical version as a base. In the end\, we illustrate the functionality of our test regarding underlying independence as well as dependence of different forms with a simulation study.
URL:https://www.hsu-hh.de/statistik/event/carina-beering-hsu-2
LOCATION:Gebäude H1\, Raum 1503
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20221012T160000
DTEND;TZID=Europe/Berlin:20221012T170000
DTSTAMP:20221005T103504Z
CREATED:20221005T103223Z
LAST-MODIFIED:20221005T103504Z
UID:975-1665590400-1665594000@www.hsu-hh.de
SUMMARY:Ángel López-Oriona (University of A Coruña\, Spain)
DESCRIPTION:Clustering of categorical time series based on two novel feature-based distances with an application to biological sequences\nTwo novel distances between categorical time series are introduced. Both of them measure discrepancy between extracted features describing the underlying serial dependence patterns. One of them is based on well-known association measures. The other relies on the so-called binarization of a categorical process\, which indicates the presence of each category by means of a canonical vector. The binarization is used to construct a collection of innovative association measures capturing every possible type of serial dependence. The metrics are used to construct crisp and fuzzy algorithms for clustering nominal series. The proposed approaches are able to group together series generated from similar underlying stochastic processes\, achieve accurate results with series coming from a broad range of models and are computationally efficient. Extensive simulation studies show that both hard and soft clustering algorithms outperform several alternative procedures presented in the literature. Two applications involving biological sequences from different species highlight the usefulness of the introduced techniques.
URL:https://www.hsu-hh.de/statistik/event/angel-lopez-oriona
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20210414T180000
DTEND;TZID=Europe/Berlin:20210414T190000
DTSTAMP:20210325T135704Z
CREATED:20210325T135552Z
LAST-MODIFIED:20210325T135704Z
UID:952-1618423200-1618426800@www.hsu-hh.de
SUMMARY:Christian Weiß (HSU)
DESCRIPTION:Soft-clipping INGARCH Models for Time Series of Bounded Counts\nThe soft-clipping binomial INGARCH (scBINGARCH) models are proposed as nearly linear time series models for bounded counts with possibly negative autocorrelations. Conditions that guarantee the existence and certain mixing properties of the scBINGARCH process are derived\, and further stochastic properties are discussed. The consistency and asymptotic normality of maximum likelihood estimators are established\, and finite-sample properties are studied with simulations. The practical relevance of the scBINGARCH’s ability to allow for negative parameter and ACF values is demonstrated by some real-data examples.
URL:https://www.hsu-hh.de/statistik/event/christian-weiss-hsu
LOCATION:BBB
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20200318T154500
DTEND;TZID=Europe/Berlin:20200318T171500
DTSTAMP:20200129T094942Z
CREATED:20200129T094804Z
LAST-MODIFIED:20200129T094942Z
UID:871-1584546300-1584551700@www.hsu-hh.de
SUMMARY:Yves Breitmoser (Uni Bielefeld)
DESCRIPTION:An axiomatic foundation of conditional logit\nThis paper considers a decision maker choosing from a set of options when options have multiple real-valued attributes. Assuming DM chooses all options with positive probability\, four invariance assumptions are necessary and sufficient for choice\nprobabilities to take McFadden’s conditional logit form: independence of irrelevant alternatives\, translation invariance\, presentation independence and context independence. Variations on these assumptions yield generalized logit and contextual logit\nmodels. This shows that even specific logit models have behavioral foundations in simple invariance assumptions involving observables only\, which therefore are\ndirectly testable.
URL:https://www.hsu-hh.de/statistik/event/an-axiomatic-foundation-of-conditional-logit
LOCATION:Gebäude H1\, Raum 1503
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20200304T154500
DTEND;TZID=Europe/Berlin:20200304T171500
DTSTAMP:20200129T120142Z
CREATED:20200129T093447Z
LAST-MODIFIED:20200129T120142Z
UID:862-1583336700-1583342100@www.hsu-hh.de
SUMMARY:Annette Möller (TU Clausthal)
DESCRIPTION:Vine copula based post-processing of ensemble forecasts for temperature\nTo account for forecast uncertainty in numerical weather prediction (NWP) models it has become common practice to employ ensemble prediction systems generating probabilistic forecast ensembles by multiple runs of the NWP\nmodel\, each time with variations in the details of the numerical model and/or initial and boundary conditions. However\, forecast ensembles typically exhibit biases and dispersion errors as they are not able to fully represent uncertainty in NWP models. Therefore\, statistical postprocessing models are employed to correct ensembles for biases and dispersion errors in conjunction with recently observed forecast errors. We propose a novel postprocessing approach for temperature forecasts based on D-vine copula quantile regression. It is a multivariate regression approach predicting quantiles of the response (temperature observations) conditioned on a set of predictor variables (the ensemble forecasts)\, while not making specific assumptions about the shape of the conditional quantiles. It exploits the dependence between observation and predictors\, accounting for non-gaussian dependencies in a flexible and data driven way. In a comparative study with temperature forecasts of different forecast horizons from the European Center for Medium Range Weather Forecast (ECMWF) the D-vine postprocessing approach shows to be highly competitive to the state-of-the-art EMOS model\, improving over standard EMOS especially for larger forecast horizons. Furthermore\, an exploratory data analysis revealed that the dependency between temperature observations and its ensemble forecasts is indeed non-Gaussian\, pointing to the need to employ vine copula models for postprocessing\, as they allow for more flexibility in the dependence structure than state-of-the-art models.
URL:https://www.hsu-hh.de/statistik/event/vine-copula-based-post-processing
LOCATION:Gebäude H1\, Raum 1503
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20200129T154500
DTEND;TZID=Europe/Berlin:20200129T171500
DTSTAMP:20200129T115746Z
CREATED:20191008T114046Z
LAST-MODIFIED:20200129T115746Z
UID:811-1580312700-1580318100@www.hsu-hh.de
SUMMARY:Houssem Brairi (USTHB Algerien)
DESCRIPTION:Testing discrete-valued time series for whiteness\nWe consider the problem of testing a univariate discrete-valued time series for whiteness in the sequency domain\, using Walsh–Fourier analysis. We show that the distribution of the lag window estimator of the Walsh spectral density is a scaled chi-square distribution\, where the scale and degrees of freedom\, both depend on the bandwidth of the smoothing window associated with the estimator. The definition of the bandwidth is extended from the frequency to the sequency domain. To address our problem\, we propose three tests: the first one is based on the cumulative Walsh periodogram\, and is shown to converge to a Brownian bridge. The second test is based on applying the Cramer–von Mises functional to an estimate of the Walsh spectral density\, and is shown to converge to a Normal distribution\, while the last test is based on a distance to whiteness\, and is shown to have an approximate scaled chi-square distribution. Simulations are reported on the performance of the tests. Finally\, we apply the proposed tests to the brain functional connectivity of schizophrenic patients.
URL:https://www.hsu-hh.de/statistik/event/testing-discrete-valued-time-series-for-whiteness
LOCATION:Gebäude H1\, Raum 1503
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20200115T080000
DTEND;TZID=Europe/Berlin:20200115T170000
DTSTAMP:20200115T172147Z
CREATED:20200115T172146Z
LAST-MODIFIED:20200115T172147Z
UID:856-1579075200-1579107600@www.hsu-hh.de
SUMMARY:Marco Meyer (TU Braunschweig)
DESCRIPTION:Extrapolation of GIDAS accident data to Germany and Europe\nWe investigate traffic accident data from project GIDAS (German In-Depth Accident Study). This project collects detailed accident data in two reference regions within Germany\, Hanover and Dresden. For each accident with severely injured persons in these regions\, the regular accident recording by police and medical staff is accompanied on site by a GIDAS team consisting of automotive engineers and accident researchers. The result is a very detailed data set that is\, however\, restricted to two particular regions. We develop methodology to extrapolate the influence of certain factors on the distribution of injury severity to other regions in Germany and Europe. The main problem is to take into account structural differences between the regions\, such as relation between traffic on highways/country roads/inner-city roads\, different vehicle types and so on. In particular\, we have to compare discrete distributions over pre-defined classes of injury severity that may vary only by small amounts across regions.
URL:https://www.hsu-hh.de/statistik/event/extrapolation-of-gidas-accident
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20191002T161500
DTEND;TZID=Europe/Berlin:20191002T174500
DTSTAMP:20190927T062601Z
CREATED:20190911T163844Z
LAST-MODIFIED:20190927T062601Z
UID:800-1570032900-1570038300@www.hsu-hh.de
SUMMARY:Alexander Schnurr (Uni Siegen)
DESCRIPTION:Ordinal Patterns and Ordinal Pattern Dependence\nOrdinal patterns describe the order structure of data points over a small time horizon. Using a moving window approach we reduce the complexity of a time series by analyzing the sequence of ordinal patterns instead of the original data. We present limit theorems for ordinal pattern probabilities and tests for structural breaks in the short-range dependent as well as in the long-range dependent setting. In the long-range dependent case\, we investigate the ordinal information of a subordinated Gaussian process with a non-summable autocovariance function. We establish the asymptotic behavior of different estimators for ordinal pattern probabilities by using a multivariate Hermite decomposition. \nOrdinal pattern dependence is a new way of measuring the degree of dependence between time series. Since it only relies on the ordinal structure of the data\, it is robust against monotone transformations and measurement errors. This method has proved to be useful already in the context of hydrological\, financial as well as medical data. Using this concept it is possible to analyze whether the dependence structure between two time series changes over time.
URL:https://www.hsu-hh.de/statistik/event/ordinal-patterns-and-ordinal
LOCATION:Gebäude H1\, Raum 1503
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190820T154500
DTEND;TZID=Europe/Berlin:20190820T154500
DTSTAMP:20190716T115847Z
CREATED:20190703T195914Z
LAST-MODIFIED:20190716T115847Z
UID:790-1566315900-1566315900@www.hsu-hh.de
SUMMARY:Burcu Aytacoglu (Ege University)
DESCRIPTION:Effect of estimation under non-normality on the phase II performance of linear profile monitoring approaches\nRecently\, there have been several studies about control charts to monitor profiles\, where the quality of a process/product is expressed as function of response and explanatory variable(s). Mostly\, it is assumed that the in-control parameter values are known and the error terms are normally distributed. However\, these assumptions are rarely satisfied in practice. In this study\, we focused on three popular methods (EWMA-R\, EWMA-3\, and EWMA3(d2)) for monitoring simple linear profiles and the performance of them is examined via simulation where the in-control parameters are estimated and error terms have a Student’s t distribution or gamma distribution. In order to capture the sampling variation among different practitioners\, average and standard deviation of the average run length (ARL) are used as performance measures. In conclusion\, it is seen that the estimation effect becomes more severe as the error term distribution deviates from normality to a greater extent. In addition\, although the average ARL values get closer to the desired values as the amount of Phase I data increases\, their standard deviations remain far away from the acceptable level indicating a high practitioner-to-practitioner variability.
URL:https://www.hsu-hh.de/statistik/event/effect-of-estimation-under-non-normality
LOCATION:Gebäude H1\, Raum 1503
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190612T163000
DTEND;TZID=Europe/Berlin:20190612T180000
DTSTAMP:20190424T123625Z
CREATED:20190424T123524Z
LAST-MODIFIED:20190424T123625Z
UID:781-1560357000-1560362400@www.hsu-hh.de
SUMMARY:Amanda Fernández-Fontelo (HU Berlin)
DESCRIPTION:INAR-hidden Markov chains to deal with misreported count data \nSince McKenzie presented the classical non-negative integer-valued autoregressive (INAR)\nmodel\, the analysis of count time series has been rapidly growing in the past years\, and many\nauthors have been actively contributing to its improvement. However\, many issues remain to\nbe addressed in this field.\nIn the present work\, the authors introduce new models of count time series that accommodate\npotential misreporting in data. Misreporting in count data is a widespread concern that is\nresponsible for reporting inaccurate levels of data. This phenomenon can appear in terms of\nunder-reporting (less than the actual amount of data) or over-reporting (more than the actual\namount of data). However\, the believability of such data significantly decreases when both\nphenomena are present.
URL:https://www.hsu-hh.de/statistik/event/inar-hidden-markov-chains-to-deal-with-misreported-count-data
LOCATION:Gebäude H1\, Raum 1503
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190612T150000
DTEND;TZID=Europe/Berlin:20190612T163000
DTSTAMP:20190424T123239Z
CREATED:20190424T123213Z
LAST-MODIFIED:20190424T123239Z
UID:777-1560351600-1560357000@www.hsu-hh.de
SUMMARY:Johannes Bracher (Uni Zürich)
DESCRIPTION:Some extensions to the endemic-epidemic model class for infectious disease surveillance counts \nThe endemic-epidemic class (Held et al 2005\, DOI 10.1191=1471082X05st098oa) is a modelling framework for multivariate infectious disease surveillance counts closely related to INGARCH models. It allows to model counts stratified by e.g. disease type\, geographical area or age group and is readily implemented in the R package surveillance. I will present some recent developments for this model class. In a first part I will talk about an extension to higher-order lags and some properties of the resulting model class\, in particular their periodically stationary properties. These can be used for model assessment and to link retrospective modelling to prospective outbreak detection. In a second part I will talk about the problem of underreporting\, which is very common in infectious disease epidemiology. I will show how inference for models with an additional reporting step can be done using an approximate maximum likelihood scheme. Moreover I will address some biases which can occur in naive analyses where underreporting is ignored. The different extensions are illustrated using various data examples.
URL:https://www.hsu-hh.de/statistik/event/some-extensions-to-the-endemic-epidemic-model
LOCATION:Gebäude H1\, Raum 1503
CATEGORIES:Kolloquium
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190529T154500
DTEND;TZID=Europe/Berlin:20190529T171500
DTSTAMP:20190416T124132Z
CREATED:20190416T124033Z
LAST-MODIFIED:20190416T124132Z
UID:767-1559144700-1559150100@www.hsu-hh.de
SUMMARY:Rainer A. Schüssler (Uni Rostock)
DESCRIPTION:Forecasting the Equity Premium: Mind the News! \nThis paper introduces a novel strategy for predicting the monthly equity premium based on extracted news from more than 700\,000 newspaper articles\, published in The New York Times and Washington Post between 1980 and 2018. We propose a flexible data-adaptive switching approach to map a large set of different news-topics into forecasts of aggregate stock returns. The information embedded in our extracted news are not captured by established equity premium predictors. Compared to the historical mean between 1999 and 2018\, we find large out-of-sample (OOS) gains with an R²oos of 6.52% and sizeable utility gains for a mean-variance investor. The empirical results imply that (geo-)political rather than economic news are more valuable to forecast the equity premium out of sample. Prediction gains arise in down markets.
URL:https://www.hsu-hh.de/statistik/event/forecasting-the-equity-premium-mind
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190522T154500
DTEND;TZID=Europe/Berlin:20190522T171500
DTSTAMP:20190416T122243Z
CREATED:20190416T122242Z
LAST-MODIFIED:20190416T122243Z
UID:755-1558539900-1558545300@www.hsu-hh.de
SUMMARY:Uwe Saint-Mont (HS Nordhausen)
DESCRIPTION:Auf der Suche nach relevanten Merkmalen \nDie Selektion relevanter Merkmale ist ein zentrales Problem der Statistik und der empirischen Wissenschaften im Allgemeinen. Zwar lassen sich heute problemlos große Datenmengen erheben\, also viele Merkmale zahlreicher statistischer Merkmalsträger festhalten\, doch welche davon sind wichtig? Wie ist das kausale Gefüge und welche Variablen steuern das Geschehen? \nSchon im 19. Jahrhundert hat John Stuart Mill diese Fragen mit einer überschaubaren Menge von Argumentationsmustern überzeugend beantwortet. Im 20. Jahrhundert hat die angewandte Statistik diese Ideen aufgegriffen und weiterentwickelt: Entweder man untersucht in einem Experiment wenige Merkmale sehr präzise oder man versucht aus einer Vielzahl von „Kandidaten“ die interessanten zu isolieren. So kommt man einerseits zu experimentellen Designs und andererseits zur statistischen Modellierung. \nBis heute sind daraus randomisierte kontrollierte Studien\, Metaanalysen\, Strukturgleichungsmodelle\, Regressionsanalysen\, epidemiologische Korrelationsstudien und kausalen Graphen hervorgegangen. So verschiedenartig die einzelnen Ansätze auch wirken: Immer ist man den Ursachen auf der Spur und wird gar nicht so selten fündig. \nAuch wenn konkrete Verfahren eine gewisse Rolle spielen werden\, so geht es im Vortrag vor allem um die wesentlichen Ideen\, also um die großen „strategischen“ Zusammenhänge.
URL:https://www.hsu-hh.de/statistik/event/auf-der-suche-nach-relevanten-merkmalen
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190424T154500
DTEND;TZID=Europe/Berlin:20190424T171500
DTSTAMP:20190416T122017Z
CREATED:20190315T113011Z
LAST-MODIFIED:20190416T122017Z
UID:748-1556120700-1556126100@www.hsu-hh.de
SUMMARY:Arne Johannssen (Uni Hamburg)
DESCRIPTION:Health Care Monitoring by Hypergeometric Control Charts for Fractions Non-Conforming \nProcess monitoring in health care organisations is one of the core tasks of an efficient medical risk management system for detecting\, assessing\, mitigating\, and preventing risks. Statistical control charts as adequate tools for process monitoring are well-suited to observe\, measure\, and improve health care outcomes as well as to minimize the occurrence of adverse events. In this presentation we focus on various improvements of control charts for „fraction non-conforming“ and their application to health care monitoring\, since (hospital) quality indicators are often binary at the patient level\, presented as proportions\, risk-adjusted or standardised rates. In particular\, we introduce a new class of statistical control charts based on the (negative) Hypergeometric distribution that comprises improvements of Binomial p-charts (with approximative or exact control limits)\, Geometric g-charts\, and negative Binomial cumulative count of conforming charts. The proposed class of Hypergeometric control charts meets numerous requirements of efficient health care monitoring and is especially useful when monitoring high-yield processes.
URL:https://www.hsu-hh.de/statistik/event/health-care-monitoring-by-hypergeometric-control
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190417T154500
DTEND;TZID=Europe/Berlin:20190417T171500
DTSTAMP:20190305T094258Z
CREATED:20190305T094258Z
LAST-MODIFIED:20190305T094258Z
UID:738-1555515900-1555521300@www.hsu-hh.de
SUMMARY:Dhouha Mejri (TU Dortmund)
DESCRIPTION:Adaptive Control charts for identifying concept drift in nonstationary environment \nTime adjusting dynamic systems whose underlying changing distribution should be continuously\nmonitored to track abnormal behaviors is one of the most recent challenges in many real life\napplications. In fields such as sensor networks\, intrusion detection\, credit card fraud detection and\nprocess monitoring\, the arriving data change over time and the target concept to be learned changes\naccordingly causing the problem of “concept drift. Adaptive control charts from SPC domain and\ndynamic ensemble methods from data mining field are the most widely used techniques to track\nconcept drift. In order to perform the change identification in data stream processes\, the first\nenhancement of ensemble methods in SPC proposed in Mejri et al\, 2018 entitled Dynamic Weighted\nMajority Control Chart will be presented. The new adaptive method has not only the ability to\ncombine more than two control charts but also uses the expertise of dynamic weighted majorityWinnow (DWM)-WIN in\nidentifying and learning changes during the monitoring. First\, it transforms\nthe task of determining the state of the process into a classification problem by treating control charts\nas attributes where the drift has to be predicted. Second\, DWM-WIN mechanism is applied to learn the\nshift and to combine the decision of different individuals. Third\, a prediction of class label is used to\nhelp in classifying the shift during the changing of the process toward the approximated right\ndirection. The three main steps of this combined control chart as well comparative results will be\npresented and discussed in this talk
URL:https://www.hsu-hh.de/statistik/event/adaptive-control-charts-for-identifying-concept
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190313T154500
DTEND;TZID=Europe/Berlin:20190313T171500
DTSTAMP:20190226T152341Z
CREATED:20190226T140207Z
LAST-MODIFIED:20190226T152341Z
UID:727-1552491900-1552497300@www.hsu-hh.de
SUMMARY:Aisouda Hoshiyar (HSU)
DESCRIPTION:Challenging the commonly used log-link in statistical models for count data with an application to infectious disease data \nA response function is an essential part of any generalized linear model\, but its choice\nis rarely questioned. In particular\, if the modeled expected value is restricted to be\ngreater than zero\, the choice often falls on the exponential function. Even for a response\nvariable\, for which the exponential function corresponds to the canonical link\, there is\nno indication that this is the true response function in general. Therefore\, we propose to\ntake the softplus function as response function into consideration. The softplus function\,\nwhich is technically used in the context of neural networks\, enables the modeling of the\nconditional mean in an additive way and therefore ensures a linear interpretation of the\nregression coefficients while respecting the positivity boundary of the conditional mean\nat the same time. The central research question to be discussed in this study is: Does\nthe softplus activating function represent an adequate substitute of the commonly used\nlog-link with an application to infectious diseases? In the first step\, a simulation study\ngives insight into the robustness of the estimated coefficients under various circumstances.\nFurthermore\, the framework for the analysis of multivariate infection disease data yield\nby Held et al. (2005) is self-implemented via the open source software R. By doing so\,\nthe softplus function is introduced to the model class applied. The estimation results\nfrom Held et al. (2005) are reproduced and compared to those concerning the softplus\nlink function with respect to the predictive quality. One-step-ahead-predictions build the\nbasis for mean-squared prediction errors and coverage frequencies of the upper prediction\nlimits. The results have been obtained using general optimisation routines via maximum\nlikelihood estimation.
URL:https://www.hsu-hh.de/statistik/event/challenging-the-commonly-used-log-link
LOCATION:Gebäude H1\, Raum 2151
CATEGORIES:Kolloquium
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190213T154500
DTEND;TZID=Europe/Berlin:20190213T171500
DTSTAMP:20190109T141313Z
CREATED:20190109T141209Z
LAST-MODIFIED:20190109T141313Z
UID:691-1550072700-1550078100@www.hsu-hh.de
SUMMARY:Annika Homburg (HSU)
DESCRIPTION:Point Forecasting in Discrete Time Series Analysis\nIn this work we determine central and non-central coherent point forecasts of various discrete valued time series models.\nEach coherent integer forecast is compared to its approximation\, derived from the model representing the continuous counterpart to each respective discrete model. Several INAR(1) processes and the influence of their distribution parameters\nare analyzed.
URL:https://www.hsu-hh.de/statistik/event/point-forecasting-in-discrete-time-series-analysis
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190206T161500
DTEND;TZID=Europe/Berlin:20190206T174500
DTSTAMP:20190128T161126Z
CREATED:20190121T162340Z
LAST-MODIFIED:20190128T161126Z
UID:704-1549469700-1549475100@www.hsu-hh.de
SUMMARY:Andreas Groll (TU Dortmund)
DESCRIPTION:Effect Selection in Cox Frailty Models by Regularization Methods\nIn all sorts of regression problems it has become more and more important to deal with complex and high dimensional data with lots of potentially influential covariates. A possible solution is to apply estimation methods that aim at the detection of the relevant effect structure by using regularization methods. In this talk\, the effect structure in the Cox frailty model\, which is the most widely used model that accounts for heterogeneity in survival data\, is investigated. Since in survival models one has to account for possible variation of the effect strength over time the selection of the relevant features has to distinguish between several cases: covariates can have time-varying effects\, can have time-constant effects or be irrelevant. A regularization approach is proposed that is able to distinguish between these types of effects to obtain a sparse representation that includes the relevant effects in a proper form. The method is applied to a real world data set\, illustrating that the complexity of the influence structure can be strongly reduced by using the proposed regularization approach.
URL:https://www.hsu-hh.de/statistik/event/effect-selection-in-cox-frailty-models-by-regularization-methods
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20190108T154500
DTEND;TZID=Europe/Berlin:20190108T171500
DTSTAMP:20181226T095130Z
CREATED:20181226T094910Z
LAST-MODIFIED:20181226T095130Z
UID:669-1546962300-1546967700@www.hsu-hh.de
SUMMARY:Maria Mohr (Uni Hamburg)
DESCRIPTION:Changepoint detection in a nonparametric time series regression model\nA weakly dependent time series is considered\, for which we develop a strategy to detect whether the nonparametric conditional mean function is stable in time. The strategy allows for autoregressive effects and heteroscedasticity. Our proposal is based on a modified CUSUM-type test procedure\, which uses a sequential marked empirical process of residuals. We show weak convergence of the considered process to a centered Gaussian process under the null ”mt(·) = m(·) for all t” and a stationarity assumption. This requires some sophisticated arguments for sequential empirical processes of weakly dependent variables. As a consequence we obtain the convergence of Kolmogorov-Smirnov\nand Cramér-von Mises type test statistics. The procedure acquires a very simple limiting distribution and nice consistency properties against changepoint alternatives\, features from which related tests are lacking. Further considerations include a bootstrap procedure as well as a test for change in the conditional variance function. Finally\, a simulation study is conducted to investigate the finite sample performance of our tests.
URL:https://www.hsu-hh.de/statistik/event/changepoint-detection-in-a-nonparametric-time
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Berlin:20181211T154500
DTEND;TZID=Europe/Berlin:20181211T171500
DTSTAMP:20181226T094847Z
CREATED:20181108T111402Z
LAST-MODIFIED:20181226T094847Z
UID:655-1544543100-1544548500@www.hsu-hh.de
SUMMARY:Tobias A. Möller (HSU)
DESCRIPTION:Integer-valued max-autoregressive models\nThe talk addresses an introduction to integer-valued max-autoregressive models. The parameter estimation for such models\, e.g.\, the max-INAR(1) model\, seems to be straightforward. The max-INAR(1) model is a Markov chain and maximum likelihood estimation with numerical optimization routines seems to be easily applicable. But if the observed counts attain very large values\, numerical issues frustrate this plan. The structure of the max-INAR(1) process will be used to show up a way to circumvent this problem. An example of the parameter estimation procedure with real data (counts of cinema visitors) demonstrates the application.
URL:https://www.hsu-hh.de/statistik/event/integer-valued-max-autoregressive-models
LOCATION:Gebäude H1\, Raum 1505\, Holstenhofweg 85\, Hamburg\, Hamburg\, 22043\, Deutschland
CATEGORIES:Kolloquium
ORGANIZER;CN="F%C3%A4chergruppe Mathematik und Statistik":MAILTO:weissc@hsu-hh.de
END:VEVENT
END:VCALENDAR