Dr. rer. nat.
Now showing 1 - 10 of 11
Publicationtanh Neurons are Bayesian Decision Makers( 2021)
; ;The hyperbolic tangent (tanh) is a traditional choice for the activation function of the neurons of an artificial neural network. Here, we go through a simple calculation that shows that this modeling choice is linked to Bayesian decision theory. Our brief, tutorial-like discussion is intended as a reference to an observation rarely mentioned in standard textbooks.
PublicationMax-Sum Dispersion via Quantum Annealing( 2019)
; ; ;We devise an Ising model for the max-sum dispersion problem which occurs in contexts such as Web search or text summarization. Given this Ising model, max-sum dispersion can be solved on adiabatic quantum computers; in proof of concept simulations, we solve the corresponding Schrödinger equations and observe our approach to work well.
PublicationA QUBO Formulation of the k-Medoids Problem( 2019)
; ; ; ;We are concerned with k-medoids clustering and propose aquadratic unconstrained binary optimization (QUBO) formulation of the problem of identifying k medoids among n data points without having to cluster the data. Given our QUBO formulation of this NP-hard problem, it should be possible to solve it on adiabatic quantum computers.
PublicationDetecting Mobility Patterns with Stationary Bluetooth Sensors: A real-world Case Study( 2015)
;Müller, Marc ; ;A Bluetooth sensor network was built up in the city of Bonn to measure Bluetooth MAC-addresses. The results of the acquired data are separated on a macro level and mobility patterns. We have collected nearly 5 million data points from 14 distinct stationary sensors over a period of 1 month and recognized over 85.000 unique devices. We show that the data is sufficiently dense to detect commuter patterns based on a Fourier analysis. In addition, we discuss limitations found in the dataset and present lessons learned.
PublicationUsing Bluetooth to track mobility patterns: Depicting its potential based on various case studies( 2013)
;Ellersiek, Timothy ;Andrienko, Gennady ;Andrienko, Natalia ; ;Müller, MarcDuring the past years the interest in the exploitation of mobility information has increased significantly. A growing number of companies and research institutions are interested in the analysis of mobility data with demand of a high level of spatial detail. Means of tracking persons in our environment can nowadays be fulfilled by utilizing several technologies, for example the Bluetooth technology, offering means to obtain movement data. This paper gives an overview of four case studies in the field of Bluetooth tracking which were conducted in order to provide helpful insights on movement aspects for decision makers in their specific microcosm. Aim is to analyse spatio-temporal validity of Bluetooth tracking, and in doing so, to describe the potential of Bluetooth in pedestrian mobility mining.
PublicationAnalyse von raumzeitlichen Bewegungsmustern auf Basis von Bluetooth-Sensoren( 2012)
;Ellersiek, Timothy ; ;Informationen über das Kundenverhalten sind ein wesentlicher Forschungsbestandteil im Bereich des Marketings. Dabei hilft das Verstehen von Entscheidungsfindungsprozessendes potenziellen Kunden in einem raumzeitlichen Wechselspiel mit seiner Umgebung dem Anbieter die Qualität seines Produktes aufzuwerten. Bis vor kurzem wurden Veränderungen am Produkt häufig durch Trial-and-Error-Methoden vorgenommen; heute jedoch erlauben neuartige Technologien wie beispielsweise Bluetooth, GPS oder Video neue Möglichkeiten, dem Anbieter zielgenaue Qualitäts-Validierungen seiner Produkte durchzuführen. In dem vorliegenden Beispiel wird der vorangehende Sachverhalt auf die Mobilität im Duisburger Zoo übertragen. Repräsentative Daten über das Bewegungsverhalten der Besucher werden unter Wahrung der Privatsphäre in einem siebentägigen Versuchszeitraum anhand von Bluetooth-Tracking erfasst, anschließend aufbereitet, analysiert und interpretiert. Die Ergebnisse zeigen, dass die Reproduktion von Mobilitätsinformationendurch die Erfassung von privaten Endgeräten der Besucher mittels einer kostengünstigen Technologie wie das Bluetooth möglich ist.
PublicationAnalytical workflow of monitoring human mobility in big event settings using bluetooth( 2011)
; ; ; ;Andrienko, GennadyAndrienko, NataliaIn recent times, consumer research at major social events received significant interest by organizing companies. Understanding the movements and motivations of the customers enables new business strategies and is needed to minimize the risk of investment. The spatiotemporal complexity of major events poses high demands on survey and analytical methods. New technological advances in both event monitoring systems and evaluation methods of movement data provide new insights into the behavioral patterns of customers by preserving their privacy. In this paper we present a work that seeks to systematize the research process of design, collection, and analysis of visitor behavior in a mixed indoor-outdoor event setting using Bluetooth sensor technology. The defined workow is comprised of 5 steps and designed to answer heterogeneous business questions with respect to customer movement behavior in a single event context. Our approach is applied in a real-world business applicati on for a Formula 1 event.
PublicationVisit potential: A common vocabulary for the analysis of entity-location interactions in mobility applications( 2010)
; ; ;A growing number of companies and public institutions use mobility data in their day-to-day business. One type of usage is the analysis of spatio-temporal interactions between mobile entities and geographic locations. In practice the employed measures depend on application demands and use context-specific terminology. Thus, a patchwork of measures has evolved which is not suitable for methodological research and interdisciplinary ex-change of ideas. The measures lack a systematic formalization and a uni-form terminology. In this paper we therefore systematically define meas-ures for entity-location interactions which we name visit potential. We provide a common vocabulary that can be applied for an entire class of mobility applications. We present two real-world scenarios which apply entity-location interaction measures and demonstrate how the employed measures can be precisely defined in terms of visit potential.
PublicationSample bias due to missing data in mobility surveys( 2010)
; ; ;A growing number of companies use mobility information in their day-to-day business. One requirement thereby is that inference about population-wide mobility patterns can be made. Therefore, it is not only important to find mobility patterns in a given data sample but also to assert their validity for the total population. This aspect of analysis has been largely neglected in mobility data mining research, which limits the applicability of the whole algorithmic field. In this paper we will analyze one aspect of sample bias due to incomplete mobility data. We will provide a systematic approach to detect dependencies between mobility behavior, socio-demography and missing data. Further, we apply the approach to a large GPS mobility survey in Switzerland and show that our concerns are justified and require attention in future research. We hope that our paper will raise the awareness that representativity of mobile behavior cannot be taken for granted in mobility surveys du e to missing data and is a research direction of utmost importance.
PublicationModelling missing values for audience measurement in outdoor advertising using GPS data( 2009)
; ; ; ;Pasquier, M. ;Hofmann, UrsMende, F.GPS technology has made it possible to evaluate the performance of outdoor advertising campaigns in an objective manner. Given the GPS trajectories of a sample of test persons over several days, their passages with arbitrary poster campaigns can be calculated. However, inference is complicated by the early dropout of persons. Other than in most demonstrations of spatial data mining algorithms where the structure of the data sample is usually disregarded, poster performance measures such as reach and gross impressions evolve continuously over time and require non-intermittent observations. In this paper, we investigate the applicability of survival analysis to compensate for missing measurement days. We formalize the task of modeling the visit potential of geographic locations based on trajectory data as our variable of interest results from dispersed events in space-time. We perform experiments on the cities of Zurich and Bern simulating different dropout mechanisms and dropout rates and show the adequacy of the applied method. Our modeling technique is at present part of a business solution for the Swiss outdoor advertising branch and serves as pricing basis for the majority of Swiss poster locations.