RESEDA is focussed on designing, developing and improving models and representational frameworks for the representation of semantic information in knowledge organization systems (KOS). We aim at developing strategies for precisely specifying the semantic content and characteristics of concepts and the semantic relations between these concepts in indexing languages and KOS, thereby augmenting the semantic richness and expressivity of these vocabularies. The work on the theoretical foundations and development of models and frameworks will be complemented by the construction of application-oriented, functional inventories and vocabularies for a large variety of applications from different fields.
An additional objective of the project is the construction of an autonomously structured KOS which serves as an ontological spine and functional backbone connecting several highly heterogeneous vocabularies and conceptual structures in order to facilitate their semantically interoperable management in modern information retrieval systems.
Semantic richness and expressivity of indexing languages and KOS is of particular importance for the semantic interoperability and advanced strategies of data exchange between heterogeneous systems and aspects of localization and internationalization of established KOS. The semantic enrichment and particularly the specification of the semantic relations used for the construction of these indexing languages are of considerable importance when designing multilingual and internationally interoperable systems. Especially applications and strategies connecting several structurally different systems and providing multilingual access shall profit considerably from advanced semantics and improved relational structures.
The development of new approaches to representational models is based on a thorough analysis of existing classification systems and verbal indexing languages and established strategies for their construction. The transfer and adaptation of modelling methods from the field of semantic networks and the selective application of various techniques for the construction, refinement of indexing languages is designed to create indexing languages and representational methods which can be used as expressive and functional ontologies.
Modern information retrieval strategies and technically advanced KOS require the detailed and comprehensive representation of semantic information in expressive and standardized formats.
RESEDA is designed as a continuation and extension of various projects in the field of localization and Internationalization of KOS and the development of projects connecting several different reference vocabularies from different languages thus allowing for a structured and functional access to multilingual and heterogeneous indexing languages and the respective repositories.
Achieving interoperability of two complex, fully developed and typologically different systems like the German reference vocabulary "Schlagwortnormdatei" (SWD) and the "Dewey Decimal Classification" (DDC) by extensive and sophisticated mapping strategies (CrissCross) can be enhanced even further by analysing and specifying the semantic relations holding between the elements of the respective systems and between the terms and their respective counterparts in each systems.
Based on results and observations from various other projects and research done in this field, RESEDA is focussed on developing and implementing new models for the organization and representation of semantic relations and conceptual interrelations in different knowledge structures.
Laying the fundaments for such a work by providing a general theory of semantic relations is another important aspect of this project.
A thorough analysis based on a synoptic view and comparison of different strategies and approaches and the idea of refining and expanding the functionalities of already established indexing languages and reference vocabularies illustrates the variety of different possible approaches. Selectively applying individual principles and techniques from the field of semantic networks combining various advantages and functionalities of this type of models in order to express and define semantic and logical information of the elements of both verbal and classificatory indexing languages and description frameworks.
Particularly when enriching and restructuring already existing verbal indexing languages the refinement of the relational inventory is crucial. Therefore methods and strategies for developing limited and structured inventories of semantic relations is a primary objective of this project. One approach in doing so is the development and generation of such inventories on the basis of the existing vocabularies or specific parts of them, by strategies of abstraction and generalization.
The thorough analysis of established strategies and approaches and their underlying methodologies shall result in a further advancement and refinement of models for the representation of semantic data in the context of the semantic web.