Evaluation of the previously implemented strategies of text-to-hypertext conversion

By means of a user study, we want to examine if the  HyTex demo prototype developed in phase I is really a better work environment and therefore facilitates the user better.

Development of topically motivated strategies of text-to-hypertext conversion

Furthermore, we examine in which ways methods of lexical chaining can be applied to German specialized texts. We are therefore currently implementing a lexical chainer which—in contrast to classic approaches for the English—is emphasized by several steps of (pre-)processing. The lexical chainer will access various semantic resources. We are especially interested in the question, by means of which extensions can we achieve a transition between lexical and topical chaining.

The following publication covers the results of an annotation pre-study: Web-based Annotation of Anaphoric Relations and Lexical Chains (project publication with our partners in Bielefeld; PDF 163 KB)

Extension of the terminological and lexical resources

In order to combine the  terminological resources created in phase I with other word nets and domain ontologies, we want to transmit them into  OWL . In addition, we are interested in the ways in which terminological word nets may automatically be extended. We hope to be able to profit from the resources evolving from this automatic extension in regard to our work concerning topically motivated strategies of text-to-hypertext conversion.

Automation of pre-processing for automatic strategies of text-to-hypertext conversion

Our pilot study on automatic annotation of definitional text segments has already supplied interesting results and search patterns. Based on these results, we want to examine how these patterns may be employed for the (fully) automatic annotation of definitional text segments. Additionally, we plan to use these results as a basis for the automatic reconstruction and annotation of topical structures.

( Deutsch )