Open Access

Gefühle beeinflussen das menschliche Verhalten, indem sie beispielsweise zu bestimmten Handlungen motivieren, vergangene Erlebnisse bewerten und die soziale Interaktion prägen. Auch bei der Aktivität der Internetsuche spielen Gefühle als subjektive Empfindungen eine wichtige Rolle, sodass sie im Fachgebiet Information Seeking Behavior erforscht werden. Die vorliegende Arbeit ist in der Disziplin der Informationswissenschaft verortet und zielt darauf ab, das Wissen über die Gefühle der Suchenden zu erweitern und daraus konstruktive Schlussfolgerungen zu ziehen. Sie geht der Frage nach, wie die Informationssuche im Internet emotional erlebt wird und welche Bedingungen und Ursachen die Suchenden als bedeutsam für ihr emotionales Erleben bei der Onlinesuche betrachten. Um dies zu erforschen, wird ein methodologischer Rahmen verwendet, der sich diesem Thema auf ganz andere Art annähert, als bisherige Forschungsarbeiten auf diesem Gebiet: Die Grounded Theory-Methodologie. Durch deren Prinzipien des Fragenstellens und Vergleichens entsteht eine Theorie, die gleichzeitig interpretierend als auch empirisch fundiert ist. Als Datengrundlage dieser Theorie dienen Leitfadeninterviews, in denen junge Erwachsene aus den USA und Deutschland ihre Eindrücke und Empfindungen bei der Internetsuche schildern. Die Teilnehmenden beziehen sich dabei auf eine unmittelbar vor dem Interview durchgeführte Internetsuche, in der sie durch ein eigenes Informationsbedürfnis angeleitet wurden. Als Ergebnis der Studie zeigt sich zum einen, wie stark die individuellen Suchthemen die Gefühle der Suchenden beeinflussen. Zum anderen ergibt die Untersuchung, dass diejenigen Gefühle, die sich auf die Ausführung der Suche beziehen, erstaunlich gering ausgeprägt sind, denn die Internetsuche wird als normale Routinehandlung empfunden. Aufgrund dieser Erkenntnisse zur Individualität und Alltäglichkeit der Sucherfahrung formuliert die vorliegende Arbeit Vorschläge für eine bessere Unterstützung der Suchenden und für die zukünftige Erforschung der affektiven Ebene bei der Onlinesuche.

In this paper, we describe our system developed for the GErman SenTiment AnaLysis shared Task (GESTALT) for participation in the Maintask 2: Subjective Phrase and Aspect Extraction from Product Reviews. We present a tool, which identifies subjective and aspect phrases in German product reviews. For the recognition of subjective phrases, we pursue a lexicon-based approach. For the extraction of aspect phrases from the reviews, we consider two possible ways: Besides the subjectivity and aspect look-up, we also implemented a method to establish which subjective phrase belongs to which aspect. The system achieves better results for the recognition of aspect phrases than for the subjective identification.

We present the German Sentiment Analysis Shared Task (GESTALT) which consists of two main tasks: Source, Subjective Expression and Target Extraction from Political Speeches (STEPS) and Subjective Phrase and Aspect Extraction from Product Reviews (StAR). Both tasks focused on fine-grained sentiment analysis, extracting aspects and targets with their associated subjective expressions in the German language. STEPS focused on political discussions from a corpus of speeches in the Swiss parliament. StAR fostered the analysis of product reviews as they are available from the website Amazon.de. Each shared task led to one participating submission, providing baselines for future editions of this task and highlighting specific challenges. The shared task homepage can be found at https://sites.google.com/site/iggsasharedtask/.

This 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.

In this paper, we investigate a semi- supervised learning approach based on neu- ral networks for nested named entity recog- nition on the GermEval 2014 dataset. The dataset consists of triples of a word, a named entity associated with that word in the first-level and one in the second-level. Additionally, the tag distribution is highly skewed, that is, the number of occurrences of certain types of tags is too small. Hence, we present a unified neural network archi- tecture to deal with named entities in both levels simultaneously and to improve gen- eralization performance on the classes that have a small number of labelled examples.

This paper presents the BECREATIVE Named Entity Recognition system and its participation at the GermEval 2014 Named Entity Recognition Shared Task (Benikova et al., 2014a). BECREATIVE uses a hybrid approach of two commonly used procedural methods, namely list-based lookups and machine learning (Naive Bayes Classification), which centers around the classifier. BECREATIVE currently reaches an F-score of 37.34 on the strict evaluation setting applied on the development set provided by GermEval.

This 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.

Collobert et al. (2011) showed that deep neural network architectures achieve state- of-the-art performance in many fundamental NLP tasks, including Named Entity Recognition (NER). However, results were only reported for English. This paper reports on experiments for German Named Entity Recognition, using the data from the GermEval 2014 shared task on NER. Our system achieves an F1 -measure of 75.09% according to the official metric.

This paper describes the GermEval 2014 Named Entity Recognition (NER) Shared Task workshop at KONVENS. It provides background information on the motivation of this task, the data-set, the evaluation method, and an overview of the participating systems, followed by a discussion of their results. In contrast to previous NER tasks, the GermEval 2014 edition uses an extended tagset to account for derivatives of names and tokens that contain name parts. Further, nested named entities had to be predicted, i.e. names that contain other names. The eleven participating teams employed a wide range of techniques in their systems. The most successful systems used state-of-the- art machine learning methods, combined with some knowledge-based features in hybrid systems.

Ironic speech act detection is indispensable for automatic opinion mining. This paper presents a pattern-based approach for the detection of ironic speech acts in German Web comments. The approach is based on a multilevel annotation model. Based on a gold standard corpus with labeled ironic sentences, multilevel patterns are deter- mined according to statistical and linguis- tic analysis. The extracted patterns serve to detect ironic speech acts in a Web com- ment test corpus. Automatic detection and inter-annotator results achieved by human annotators show that the detection of ironic sentences is a challenging task. However, we show that it is possible to automatically detect ironic sentences with relatively high precision up to 63%.

We present Sentilyzer, a web-based tool that can be used to analyze and visualize the sentiment of German user comments on Facebook pages. The tool collects comments via the Facebook API and uses the TreeTagger to perform basic lemmatization. The lemmatized data is then analyzed with regard to sentiment by using the Berlin Affective Word List – Reloaded (BAWL-R), a lexicon that contains emotional valence ratings for more than 2,900 German words. The results are visualized in an interactive web interface that shows sentiment analyses for single posts, but also provides a timeline view to display trends in the sentiment ratings.

This software demonstration paper presents a project on the interactive visualization of social media data. The data presentation fuses German Twitter data and a social relation network extracted from German online news. Such fusion allows for comparative analysis of the two types of media. Our system will additionally enable users to explore relationships between named entities, and to investigate events as they develop over time. Cooperative tagging of relationships is enabled through the active involvement of users. The system is available online for a broad user audience.

Political debates bearing ideological references exist for long in our society; the last few years though the explosion of the use of the internet and the social media as communication means have boosted the production of ideological texts to unprecedented levels. This creates the need for automated processing of the text if we are interested in understanding the ideological references it contains. In this work, we propose a set of linguistic rules based on certain criteria that identify a text as bearing ideology. We codify and implement these rules as part of a Natural Language Processing System that we also present. We evaluate the system by using it to identify if ideology exists in tweets published by French politicians and discuss its performance.

We introduce two corpora gathered on the web and related to computer-mediated communication: blog posts and blog comments. In order to build such corpora, we addressed following issues: website discovery and crawling, content extraction constraints, and text quality assessment. The blogs were manually classified as to their license and content type. Our results show that it is possible to find blogs in German under Creative Commons license, and that it is possible to perform text extraction and linguistic annotation efficiently enough to allow for a comparison with more traditional text types such as newspaper corpora and subtitles. The comparison gives insights on distributional properties of the processed web texts on token and type level. For example, quantitative analysis reveals that blog posts are close to written language, while comments are slightly closer to spoken language.

It is of interest to study sentence construction for children’s writing in order to understand grammatical errors and their influence on didactic decisions. For this purpose, this paper analyses sentence structures for various age groups of children’s writings in contrast to text taken from children’s and youth literature. While valency differs little between text type and age group, sentence embellishments show some differences. Both use of adjectives and adverbs increase with age and book levels. Furthermore books show a larger use thereof. This work presents one of the steps in a larger ongoing effort to understand children’s writing and reading competences at word and sentence level. The need to look at variable from non-variable features of sentence structures separately in order to find distinctive features has been an important finding.

This 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.

This paper will have a holistic view at the field of corpus-based linguistic typology and present an overview of current advances at Leipzig University. Our goal is to use automatically created text data for a large variety of languages for quantitative typological investigations. In our approaches we utilize text corpora created for several hundred languages for cross-language quantitative studies using mathematically well-founded methods (Cysouw, 2005). These analyses include the measurement of textual characteristics. Basic requirements for the use of these parameters are also discussed. The measured values are then utilized for typological studies. Using quantitative methods, correlations of measured properties of corpora among themselves or with classical typological parameters are detected. Our work can be considered as an automatic and language-independent process chain, thus allowing extensive investigations of the various languages of the world.

The paper proposes a meta language model that can dynamically incorporate the influence of wider discourse context. The model provides a conditional probability in forms of P (text|context), where the context can be arbitrary length of text, and is used to influence the probability distribution over documents. A preliminary evaluation using a 3-gram model as the base language model shows significant reductions in perplexity by incorporating discourse context.

Challenging the assumption that traditional whitespace/punctuation-based tokenisation is the best solution for any NLP application, I propose an alternative approach to segmenting text into processable units. The proposed approach is nearly knowledge-free, in that it does not rely on language-dependent, man-made resources. The text segmentation approach is applied to the task of automated error reduction in texts with high noise. The results are compared to conventional tokenisation.

A common way to express sentiment about some product is by comparing it to a different product. The anchor for the comparison is a comparative predicate like “better”. In this work we concentrate on the annotation of multiword predicates like “more powerful”. In the single-token-based approaches which are mostly used for the automatic detection of comparisons, one of the words has to be selected as the comparative predicate. In our first experiment, we investigate the influence of this decision on the classification performance of a machine learning system and show that annotating the modifier gives better results. In the annotation conventions adopted in standard datasets for sentiment analysis, the modified adjective is annotated as the aspect of the comparison. We discuss problems with this type of annotation and propose the introduction of an additional argument type which solves the problems. In our second experiment we show that there is only a small drop in performance when adding this new argument type.

Researchers in the (digital) humanities have identified a large potential in the use of automatic text analysis capabilities in their text studies, scaling up the amount of material that can be explored and allowing for new types of questions. To support the scholars’ specific research questions, it is important to embed the text analysis capabilities in an appropriate navigation and visualization framework. However, study-specific tailoring may make it hard to migrate analytical components across projects and compare results. To overcome this issue, we present in this paper the first version of TEANLIS (text analysis for literary scholars), a flexible framework designed to include text analysis capabilities for literary scholars.

We 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.

We present a set of refined categories of interoperability aspects and argue that the representational aspect of interoperability and its content-related aspects should be treated independently. While the implementation of a generic exchange format provides for representational interoperability, content-related interoperability is much harder to achieve. Applying a task-based approach to content-related interoperability reduces complexity and even allows for the combination of very different resources.

We study the influence of information structure on the salience of subjective expressions for human readers. Using an online survey tool, we conducted an experiment in which we asked users to rate main and relative clauses that contained either a single positive or negative or a neutral adjective. The statistical analysis of the data shows that subjective expressions are more prominent in main clauses where they are asserted than in relative clauses where they are presupposed. A corpus study suggests that speakers are sensitive to this differential salience in their production of subjective expressions.

Coreferences 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.