University of Hildesheim
Not a member yet
1144 research outputs found
Sort by
Mining corpora of computer-mediated communication: Analysis of linguistic features in Wikipedia talk pages using machine learning methods
No full textMachine learning methods offer a great potential to automatically investigate large amounts of data in the humanities. Our contribution to the workshop reports about ongoing work in the BMBF project KobRA (http://www.kobra.tu-dortmund.de) where we apply machine learning methods to the analysis of big corpora in language-focused research of computer-mediated communication (CMC). At the workshop, we will discuss first results from training a Support Vector Machine (SVM) for the classification of selected linguistic features in talk pages of the German Wikipedia corpus in DeReKo provided by the IDS Mannheim. We will investigate different representations of the data to integrate complex syntactic and semantic information for the SVM. The results shall foster both corpus-based research of CMC and the annotation of linguistic features in CMC corpora
HATNER: Nested Named Entitiy Recognition for German
No full textThis paper describes our classification and rule-based attempt at nested Named Entity Recognition for German. We explain how both approaches interact with each other and the resources we used to achieve our results. Finally, we evaluate the overall performance of our system which achieves an F-score of 52.65% on the development set and 52.11% on the final test set of the GermEval 2014 Shared Task
Brandenburg, Schlesien und Ungarn im Glogauer Erbfolgestreit (1476‒1482). Mit einer Einführung in die ungarische Forschung von Katalin Wittmann
No full textDarstellung des Krieges und der diplomatischen Initiativen während des Glogauer Erbfolgestreits (1476-1482) zwischen dem Kurfürstentum Brandenburg, dem Herzogtum Sagan sowie den Königreichen Ungarn und Böhmen
Enforcing Consistent Translation of German Compound Coreferences
No full textCoreferences to a German compound (e.g. Nordwand) can be made using its last constituent (e.g. Wand). Intuitively, both coreferences and the last constituent of the compound should share the same translation. However, since Statistical Machine Translation (SMT) systems translate at sentence level, they both may be translated inconsistently across the document. Several studies focus on document level consistency, but mostly in general terms. This paper presents a method to enforce consistency in this particular case. Using two in-domain phrase-based SMT systems, we analyse the effects of compound coreference translation consistency on translation quality and readability of documents. Experimental results show that our method improves correctness and consistency of those coreferences as well as document readability
Fast Optimal Path Search on Digital Road Maps -- Development, Implementation, and Example of Use at a Logistics Service Provider
No full textThe solution to many real-world transportation planning problems is combinatorially comprehensive, complex, and some of the necessary information is not available in a sufficiently structured form. Hence, standard solving methods are often not applicable. This holds for courier and express providers' operational dispatching, where vehicles are assigned to customer orders. Typically, dialog-oriented decision support systems (DSS) are used to generate recommendations from which a human dispatcher selects the most profitable one by means of domain-specific knowledge. In this process, solutions that consolidate the freight of multiple customer orders onto a single vehicle are usually particularly favorable. This work examines optimization of the dispatching process with particular attention to the potential of freight consolidation. In order to uncover possible savings by consolidating, many alternative vehicle routes need to be determined and compared with each other. For this purpose, fast routing algorithms are essential. These algorithms form the first focus of this work. Based on a comparative literature overview of classical'' as well as current routing algorithms, the center of attention is put on contraction hierarchies (CH) as they provide a particularly favorable ratio of query speed-up to additional memory requirement. Both optimizations of hierarchy creation and new enhancements for queries in CH are studied. The considerable merits of these approaches are empirically shown through extensive benchmarks on digital road maps provided by the OpenStreetMap project. More realistic routes are obtained when routing algorithms take turn restrictions into account. First, common approaches from literature are presented for this purpose. Next, the adaptive search is introduced as a more efficient new approach, and adjusted for usage in CH. The above-mentioned query enhancements can easily be applied to this method. The advantages of the adaptive search in CH over the edge-based search that is employed for turn restriction aware routing in earlier literature is emphasized by further benchmarks. For the second focus of this work --~the vehicle routing problem with freight consolidation~-- first, a thorough mathematical model is developed. Next, model assumptions are extracted from past operational real-world data provided by IN tIME Express Logistik GmbH. Eventually, the recursive savings algorithm is introduced as a heuristic that can generate freight consolidation recommendations in the context of a DSS. Employed in multiple benchmarks based upon the past operational data, it reveals considerable cost savings compared to former dispatching decisions.Die Lösung vieler straßengebundener Transportplanungsprobleme aus der Praxis ist kombinatorisch aufwendig, komplex und manche der notwendigen Informationen liegen nicht hinreichend strukturiert vor. Standard-Lösungsverfahren sind daher für sie oft ungeeignet. Dies gilt auch für die operative Disposition bei Express- und Direkt-Kurierdienstleistern, bei der Entscheidungen über die Zuweisung von Fahrzeugen zu Aufträgen getroffen werden. Üblicherweise kommen für diese Aufgabe dialogorientierte Entscheidungsunterstützungssysteme (EUS) zum Einsatz, welche Vorschläge generieren, aus denen menschliche Disponenten unter Einbezug von domänenspezifischem Wissen den günstigsten auswählen. Lösungen, bei denen die Frachten mehrerer Aufträge auf einem Fahrzeug konsolidiert werden, sind dabei häufig besonders wirtschaftlich. Diese Arbeit beschäftigt sich mit der Optimierung der Disposition im Hinblick auf das Konsolidierungspotential der Frachten. Zum Aufdecken möglicher Ersparnisse durch Konsolidierung müssen viele alternative Fahrzeugrouten ermittelt und verglichen werden. Hierfür sind schnelle Verfahren zur Wegsuche von zentraler Bedeutung. Diese Verfahren bilden den ersten Schwerpunkt der Arbeit. Aufbauend auf einem vergleichenden Überblick "klassischer"' sowie aktueller Wegsucheverfahren wird mit den Contraction Hierarchies (CH) ein hierarchisches Verfahren in den Mittelpunkt gestellt, das ein besonders vorteilhaftes Verhältnis von Suchgeschwindigkeit zu benötigtem Speicherbedarf aufweist. Untersucht werden Optimierungen der Hierarchie-Erzeugung sowie neue Erweiterungen der Wegsuche in CH. In ausführlichen Benchmarks auf digitalen Straßenkarten des OpenStreetMap-Projekts werden die deutlichen Verbesserungen durch diese Erweiterungen empirisch nachgewiesen. Eine größere Realitätsnähe der berechneten Routen ergibt sich durch die Berücksichtigung von Abbiegebeschränkungen bei der Wegsuche. Nach der Vorstellung der in der Literatur hierfür üblichen Ansätze wird mit der adaptiven Wegsuche ein leistungsstarkes neues Verfahren für diesen Zweck präsentiert und für den Einsatz in CH angepasst. Die oben genannten Erweiterungen der Wegsuche sind mühelos auf dieses Verfahren übertragbar. Weitere Benchmarks unterstreichen die Vorteile der adaptiven gegenüber der in früherer Literatur eingesetzten pfeilbasierten Suche nach Wegen mit Abbiegebeschränkungen in CH. Für den zweiten Schwerpunkt der Arbeit, das Konsolidierungsproblem, wird zunächst ein ausführliches mathematisches Modell entwickelt. Es folgt eine Extraktion realitätsnaher Modellannahmen aus operativen Vergangenheitsdaten der IN tIME Express Logistik GmbH. Mit dem rekursiven Savingsverfahren wird schließlich eine neue Heuristik präsentiert, die es ermöglicht, Konsolidierungsvorschläge im Rahmen eines EUS zu generieren. Sie offenbart in mehreren, auf den Vergangenheitsdaten basierenden Benchmarks ein deutliches Kosteneinsparungspotential gegenüber früheren Dispositionsentscheidungen
Centering Theory in natural text: a large-scale corpus study
No full textWe present an extensive corpus study of Centering Theory (CT), examining how adequately CT models coherence in a large body of natural text. A novel analysis of transition bigrams provides strong empirical support for several CT-related linguistic claims which so far have been investigated only on various small data sets. The study also reveals genre-based differences in texts’ degrees of entity coherence. Previous work has shown unsupervised CT-based coherence metrics to be unable to outperform a simple baseline. We identify two reasons: 1) these metrics assume that some transition types are more coherent and that they occur more frequently than others, but in our corpus the latter is not the case; and 2) the original sentence order of a document and a random permutation of its sentences differ mostly in the fraction of entity-sharing sentence pairs, exactly the factor measured by the baseline
Verb Polarity Frames: a New Resource and its Application in Target-specific Polarity Classification
No full textWe discuss target-specific polarity classification for German news texts. Novel, verb-specific features are used in a Simple Logistic Regression model. The polar perspective a verb casts on its grammatical roles is exploited. Also, an additional, largely neglected polarity class is examined: controversial texts. We found that the straightforward definition of ’controversial’ is problematic. More or less balanced polarities in a text are a poor indicator of controversy. Instead, non-polar wording helps more than polarity aggregation. However, our novel features proved useful for the remaining polarity classes
Atomic: an open-source software platform for multi-level corpus annotation
No full textThis paper presents Atomic, an open-source platform-independent desktop application for multi-level corpus annotation. Atomic aims at providing the linguistic community with a user-friendly annotation tool and sustainable platform through its focus on extensibility, a generic data model, and compatibility with existing linguistic formats. It is implemented on top of the Eclipse Rich Client Platform, a pluggable Java-based framework for creating client applications. Atomic - as a set of plug-ins for this framework - integrates with the platform and allows other researchers to develop and integrate further extensions to the software as needed. The generic graph-based meta model Salt serves as Atomic’s domain model and allows for unlimited annotation levels and types. Salt is also used as an intermediate model in the Pepper framework for conversion of linguistic data, which is fully integrated into Atomic, making the latter compatible with a wide range of linguistic formats. Atomic provides tools for both less experienced and expert annotators: graphical, mouse-driven editors and a command-line data manipulation language for rapid annotation
Named Entity Recognition for German Using Conditional Random Fields and Linguistic Resources
No full textThis paper presents a Named Entity Recognition system for German based on Conditional Random Fields. The model also includes language-independant features and features computed form large coverage lexical resources. Along side the results themselves, we show that by adding linguistic resources to a probabilistic model, the results improve significantly
Speak Your Mind, but Watch Your Mouth: Complaints in CouchSurfing References
No full textThe authors discuss how mutual criticism is expressed in the CouchSurfing community. As this community is based on mutual trust and the willingness to provide overnight accommodation in their own homes, user ratings that contain criticism and negative judgement have to be formulated in a way to avoid further conflicts and to maintain a good host image. This is why many negative evaluations contain mitigating strategies that anticipate future interactions in the community and that can be judged as face work