Alexander Mehler

Web-based Annotation of Anaphoric Relations and Lexical Chains

Annotating large text corpora is a time-consuming effort. Although single-user annotation tools are available, web-based annotation applications allow for distributed annotation and file access from different locations. In this paper we present the web-based annotation application Serengeti for annotating anaphoric relations which will be extended for the annotation of lexical chains.

STRUCTURAL SIMILARITIES OF COMPLEX NETWORKS: A COMPUTATIONAL MODEL BY EXAMPLE OF WIKI GRAPHS

This article elaborates a framework for representing and classifying large complex networks by example of wiki graphs. By means of this framework we reliably measure the similarity of document, agent, and word networks by solely regarding their topology. In doing so, the article departs from classical approaches to complex network theory which focuses on topological characteristics in order to check their small world property. This does not only include characteristics that have been studied in complex network theory, but also some of those which were invented in social network analysis and hypertext theory. We show that network classifications come into reach which go beyond the hypertext structures traditionally analyzed in web mining. The reason is that we focus on networks as a whole as units to be classified—above the level of websites and their constitutive pages. As a consequence, we bridge classical approaches to text and web mining on the one hand and complex network theory on the other hand. Last but not least, this approach also provides a framework for quantifying the linguistic notion of intertextuality.

Automatic Language Classification by means of Syntactic Dependency Networks

Abstract This article presents an approach to automatic language classification by means of linguistic networks. Networks of 11 languages were constructed from dependency treebanks, and the topology of these networks serves as input to the classification algorithm. The results match the genealogical similarities of these languages. In addition, we test two alternative approaches to automatic language classification – one based on n-grams and the other on quantitative typological indices. All three methods show good results in identifying genealogical groups. Beyond genetic similarities, network features (and feature combinations) offer a new source of typological information about languages. This information can contribute to a better understanding of the interplay of single linguistic phenomena observed in language.

Towards Structure-sensitive Hypertext Categorization

Hypertext categorization is the task of automatically assigning category labels to hypertext units. Comparable to text categorization it stays in the area of function learning based on the bag-of-features approach. This scenario faces the problem of a many-to-many relation between websites and their hidden logical document structure. The paper argues that this relation is a prevalent characteristic which interferes any effort of applying the classical apparatus of categorization to web genres. This is confirmed by a threefold experiment in hypertext categorization. In order to outline a solution to this problem, the paper sketches an alternative method of unsupervised learning which aims at bridging the gap between statistical and structural pattern recognition (Bunke et al. 2001) in the area of web mining.

Social Semantics and Its Evaluation by Means of Semantic Relatedness and Open Topic Models

This paper presents an approach using social semantics for the task of topic labelling by means of Open Topic Models. Our approach utilizes a social ontology to create an alignment of documents within a social network. Comprised category information is used to compute a topic generalization. We propose a feature-frequency-based method for measuring semantic relatedness which is needed in order to reduce the number of document features for the task of topic labelling. This method is evaluated against multiple human judgement experiments comprising two languages and three different resources. Overall the results show that social ontologies provide a rich source of terminological knowledge. The performance of the semantic relatedness measure with correlation values of up to .77 are quite promising. Results on the topic labelling experiment show, with an accuracy of up to .79, that our approach can be a valuable method for various NLP applications.

Web corpus mining by instance of Wikipedia

In this paper we present an approach to structure learning in the area of web documents. This is done in order to approach the goal of webgenre tagging in the area of web corpus linguistics. A central outcome of the paper is that purely structure oriented approaches to web document classification provide an information gain which may be utilized in combined approaches of web content and structure analysis.

Geography of social ontologies: Testing a variant of the Sapir-Whorf Hypothesis in the context of Wikipedia

In this article, we test a variant of the Sapir-Whorf Hypothesis in the area of complex network theory. This is done by analyzing social ontologies as a new resource for automatic language classification. Our method is to solely explore structural features of social ontologies in order to predict family resemblances of languages used by the corresponding communities to build these ontologies. This approach is based on a reformulation of the Sapir-Whorf Hypothesis in terms of distributed cognition. Starting from a corpus of 160 Wikipedia-based social ontologies, we test our variant of the Sapir-Whorf Hypothesis by several experiments, and find out that we outperform the corresponding baselines. All in all, the article develops an approach to classify linguistic networks of tens of thousands of vertices by exploring a small range of mathematically well-established topological indices.

Enhancing document modeling by means of open topic models Crossing the frontier of classification schemes in digital libraries by example of the DDC

Purpose – The purpose of this paper is to present a topic classification model using the Dewey Decimal Classification (DDC) as the target scheme. This is to be done by exploring metadata as provided by the Open Archives Initiative (OAI) to derive document snippets as minimal document representations. The reason is to reduce the effort of document processing in digital libraries. Further, the paper seeks to perform feature selection and extension by means of social ontologies and related web‐based lexical resources. This is done to provide reliable topic‐related classifications while circumventing the problem of data sparseness. Finally, the paper aims to evaluate the model by means of two language‐specific corpora. The paper bridges digital libraries, on the one hand, and computational linguistics, on the other. The aim is to make accessible computational linguistic methods to provide thematic classifications in digital libraries based on closed topic models such as the DDC.Design/methodology/approach – T...

Towards logical hypertext structure

Facing the retrieval problem according to the overwhelming set of documents online the adaptation of text categorization to web units has recently been pushed. The aim is to utilize categories of web sites and pages as an additional retrieval criterion. In this context, the bag-of-words model has been utilized just as HTML tags and link structures. In spite of promising results this adaptation stays in the framework of IR specific models since it neglects the content-based structuring inherent to hypertext units. This paper approaches hypertext modelling from the perspective of graph-theory. It presents an XML-based format for representing websites as hypergraphs. These hypergraphs are used to shed light on the relation of hypertext structure types and their web-based instances. We place emphasis on two characteristics of this relation: In terms of realizational ambiguity we speak of functional equivalents to the manifestation of the same structure type. In terms of polymorphism we speak of a single web unit which manifests different structure types. It is shown that polymorphism is a prevalent characteristic of web-based units. This is done by means of a categorization experiment which analyses a corpus of hypergraphs representing the structure and content of pages of conference websites. On this background we plead for a revision of text representation models by means of hypergraphs which are sensitive to the manifold structuring of web documents.

TOWARDS A SIMULATION MODEL OF DIALOGICAL ALIGNMENT

This paper presents a model of lexical alignment in communication. The aim is to provide a reference model for simulating dialogs in naming game-related simulations of language evolution. We introduce a network model of alignment to shed light on the law-like dynamics of dialogs in contrast to their random counterpart. That way, the paper provides evidence on alignment to be used as reference data in building simulation models of dyadic conversations.