Today research data and output from nearly every disciplinary and interdisciplinary domain exist also (and sometimes only) in digital form. Consequently, to ensure the data's e cient retrieval as well as long term usage and relevance, Knowledge Organisation and its tools/systems, i.e. lexical resources like thesauri and ontologies, play a major role in the Digital Libraries world. In this paper we discuss our approach of pursuing the conversion and mapping of two of the most promising lexical resources, the Information Coding Classi cation and the MultiWordNet, after having converted them into the EuroWordNet RDF/OWL format. In a second step we show how to integrate this additional knowledge into a domain Knowledge Organisation System (KOS). In the end we will have a presentation of the method and mapping as well as a use case for its evaluation for Information Retrieval purposes.
1.1
The relevance of Knowledge Organisation has gained in importance with the emergence of the digital era and the World Wide Web. Just as in the physical world these large arising digital libraries require adequate and thorough structuring systems in order to organise the increasing amount of available data and to avoid loss of information. Making them accessible by "putting them into the Semantic Web" requires thorough description and homogenisation, i.e. the provision of a data model capable of dealing with large amounts of data from various domains.
"A KOS serves as a bridge between the user's information need and the material in the collection. With it, the user should be able to identify an object of interest without prior knowledge of its existence."1 1 https://www.clir.org/pubs/reports/pub91/1knowledge.html
The Georg Eckert Institute2 (GEI) deals with various Textbook Research resources from multiple disciplines. Research data, produced within in a speci c context under particular research questions, results a broad spectrum of digital resources, e.g. a digital collection of historic textbooks3, repositories for curricula4 and textbook research related publications5, as well as digital information services6.
Because there still does not exist a proper Knowledge Management at the GEI as would be provided by using a suitable KOS these research results can not yet be linked to other research contexts and therefore are in danger of not being used afterwards. Their production within a highly specialized research community with complex but separate contexts and systems prevents them from being found easily, consequently followed by death of data", double work and waste of resources. Another researcher with related information interests has no knowledge about the already existing GEI data and is not able to satisfy their need of information easily.
Although many data models and classi cation systems have been developed
to help here, their inadequacy is evident, as these models, too, have their
intrinsic structures and characteristics and are most often bound to their respective
domain. Digital Libraries su er from too specialised models, which are not able
to communicate with each other. The advantages of the Linked Data World
ing Classi cation (ICC)
The initial development of the ICC as a "faceted classi cation" by Dahlberg
in the early seventies was made possible through a funding by the former Dt.
Gesellschaft fr Dokumentation e.V. (DGD)7. Dahlberg's intention was to
establish a classi cation system not depending on the division into knowledge
disciplines in order to facilitate long term usage and extensibility
2014, 4) pursued the ndings in the context of improving Information Retrieval
results by means of NLP techniques. He decided to take the ICC into
consideration in his approach of extending the classi cation resources WordNet Domains
(DGI). http://dgi-info.de/
and attributes). An example is shown in gure 1. The conversion supports
standardisation and hence the long term usage, as well as the EuroWordNet's
multilingualism. In
The work's continuation presented in this paper comprises several steps: the
rst step will be the actualisation and completion of the ICC itself, namely the
integration of the fourth to sixth level. This will be achieved by completing
Dahlberg's preliminary project "Logstruktur"
This part applies the conversion method from
As a proof of concept, we decided to add a third step to link the restoration domain ontology intended for document indexing purposes: the chosen restoration terminologies have been established especially for the SemRes project and used only on an experimental level.
The integration is based on the approach in
Eventually we will be able to perform Information Retrieval tasks with the converted KOS, which will test the newly integrated ICC top ontology for Information Retrieval purposes. We imagine use cases, in which Information Retrieval is done in the restoration domain, searching for the semantic concept "democracy". Through the newly established links between restoration domain ontology and the ICC Knowledge Domains via MultiWordNet, information on the "democracy" concept may also be found in the textbook research domain, where the concept would be used in the context of education or art history. This way, abstraction of concepts, integrated into the available Lexical Linked Data Cloud, will be facilitated, and the discovery of relations between di erent domains will be allowed. 3
In this paper we have shown our approach how to establish a "Lexikon der Wissensgebiete" by pursuing the work of Dahlberg (1974, 1977, 1979). We then discussed how to bring this knowledge database into the lexical linked data cloud by converting it into RDF/OWL to use it as a Top Ontology within the Lexical Resource WordNet. This step is based on the EuroWordNet conversion into RDF/OWL, presented by De Luca et al. (2007). By integrating the restoration 8 "After having found the correct SynSet, we merged manually the complete converted domain ontology under the appropriate hyperonym (SynSet). In this case we could enlarge the EuroWordNet coverage with domain-speci c terms." De Luca et al. 2007, 9. 9 "The rst step before including the domain ontologies in the new EuroWordNet OWL hierarchy was to convert these in the same OWL format [] merging methods for including these domain-ontologies to the EuroWordNet Owl representation [] domain ontology is then added to the generic one, directly under its new hyperonym." De Luca et al. 2007, 9. domain ontology as a use case we can evaluate the usefulness of the enriched lexical resource for Information Retrieval or Query expansion tasks.
Further possible work comprises the connection of additional domain ontologies, e.g. archaeology, to evaluate the usefulness of the produced and enriched lexical linked data knowledge repository in Digital Humanities research scenarios. We imagine this to o er a general improvement for Information Retrieval within the Digital Libraries world.