Ontology development from the encyclopaedic organization of knowledge

Dević, Ivana. (2017). Ontology development from the encyclopaedic organization of knowledge. PhD Thesis. Filozofski fakultet u Zagrebu, Department of Information Science
Department of Information Science > Chair of natural language processing, lexicography and encyclopedic science.
(Poslijediplomski doktorski studij informacijskih i komunikacijskih znanosti) [mentor Boras, Damir and Jecić, Zdenko].

Preview

PDF (Croatian) Download (7MB) | Preview

Abstract

The global network, the Internet, is hurriedly tranforming into an semantic network by turning from document connection onto data connection, i.e. today's WEB portals with classic information databases are becoming linked data of global computing. By comparing organisational structures of traditional encyclopedias on paper media with electronical encyclopedias in Web surroundings, you can notice certain differences that come from different types of media which enable new functionalities of searching. The above mentioned changes, through which an encyclopedical work passes through, demands an establishment of a new way of modeling organisation of encyclopedical knowledge in a Web surrounding that will find its foundation on the analysis of the specificity of a structure of an encyclopedical chapter by respecting the basic settings of the Semantic Web, principles of ontological shaping and needs of users so their availability and usability would be ensured. With the aim of preserving encyclopedical usability in today's Web information explosion, there is a need of modifying presentation of its content in a meaningful (semantic way) in a Web surrounding. Semantic interoperability means an existence of infrastructure which will enable mechanical interpretation and conclusion about content on the Web. Therefore, the key term of Semantic Web is presented by ontology, the basic component in enabling semantic interoperability. The aim of this doctoral thesis is to find out which elements of encyclopedical knowledge organisation can offer support for the development of ontology and develop a method by which ontology will be generated on the basis of encyclopedically organised knowledge. Developed ontology will be used for organisation, searching and browsing data of Croatia's Web Encyclopedia in the selected field of Croatian literature, as well as receiving precise answers on asked questions. In the introductory part of this work are presented its starting points, goals and methods, as well as the structure of the entire work. The second chapter deals with a theoretical display and clarification os Semantic Web with the aim of its full understanding. The goal of this chapter is to point out the basic theoretical and technical background of Semantic Web, the meaning of the term Semantic Web is explained, basic difference between the Web we know today and its development toward Semantic Web, disadvantages of today's Web and advantages of Semantic Web are explained, basic terms and the architecture of Semantic Web, review of basic ontologic definitions and their main goal and role. The chapter gives a detailed review of basic languages used on the semantic web with actual examples (RDF, RDFS; OWL and SKOS). Some of the more important projects of Semantic Web are shown in the third chapter of the work. During the selection of significant projects, it was considered to choose projects which are significant for better understanding of Semantic Web in the field of encyclopedics. That is why one part of this chapter deals with showing ontologies unavoidable for better understanding of Semantic Web, and the other part of the chapter gives a review of ontologic projects created entirely on encyclopedical knowledge. The mentioned analysis of existing encyclopedical ontological projects shows that no former project did try to connect and research the development of ontology and its constructive elements on the basis of structural organisation of the encyclopedial chapter by researching significance of individual structural elements of an encyclopedical chapter for the development of ontology. The fourth chapter is an introduction to the development of the ontological model of literature and to the basic settings of Protégé software. Elements of standards and ontologic languages (i.e. vocabularies RDFS, OWL, SKOKS) are shown and applied in the development of ontology of Croatian Encyclopedia in the field of Croatian literature. Pointed out was the possibility of reaching interoperability inasmuch individual ontologic resources, as well as overall ontology, with existing semantic ontologic projects on the Semantic Web which will allow conneting relevant and rich knowledge of Croatian Encyclopedia into an „global network of knowledge“ which appears and develops through projects of Semantic Web. The fifth chapter gave an insight into the historical development of encyclopedia in the world so people could completely realize the context through which encyclopedia had to go through in other to gain today's familiar features of a modern encyclopedistic work. The chapter has given basic information about the development of the Croatian central lexicographical institution; Lexicographical institution „Miroslav Krleža“ (LZMK) that does lexicography and encyclopedics of particular interest for the Republic of Croatia. It is shown that this doctoral thesis can contribute to ,with some particular elements , realising the mission and vision of LZMK. Publishing work of basic and expert enyclopedic editions of LZMK is shown, embracing editions on a paper media and those in a web surrounding. By analysing publishing work of web editions, it was found that their substantiation was mostly a matter of adapting traditional organisations of encyclopedic knowledge from paper media to web surroundings. That is why this doctoral thesis will research which uses from applying ontologic principles of semantic web would LZMK have, as well as users of these valuable knowledge sources. The sixth chapter has shown a structural organisation of encyclopedical articles, looking back onto basic types of encyclopedic articles and especially on features of biographic encyclopedial article that is the basis of research of this work. An analysis of type of data which contain encyclopedical biographic articles has been done. The mentioned allowed establishing basic ontologic layers i.e. facets by which ontologic relations will classify. The seventh chapter identifies constructive ontologic elements from encyclopedical biographic article, shows the metodology used in development of ontology, as well as resulting conceptual taxonomy and ontologic relations. The chapter has shown the role of structural elements of encyclopedical biographic article during ontologic development and connected them to corresponding constructive ontologic elements. The results of the research are presented through display of ontologic modules and belonging ontologic relations that can be used in describing a certain term. A display of structure and sequence of elements of encylopedical biographic article has been shown, with developed ontologic features by which can all types of information in the article be stored, from which constant elements have been determined that can be perceived for development of article infoframes, which is suitable for a quick insight into most important information of individual articles. Final results are shown through application of the gained ontology in description of an individual, as well as through possibility of installing complex semantic questions by unstructured data of encyclopedical biographic articles and through the possibility of organising browsing encyclopedical data, which hasn't been possible until now. The eight chapter explained the FCA approach applied during building of ontology so a gathering of conceptual features would be established, by which terms in ontology were defined so classification of terms could be made into a hierarchy. Important definitions were emphasized to understand places of formal term analysis in methodology of creating ontology. An actual example of accomplishing FCA approach was shown in 37 articles of Croatian literature in Croatian Encyclopedia, as well as a transformation of a transformational grid into a formal language of first order logics. Chapter shows the advantages of applying FCA analysis because of generating new and unfamilliar terms which could be hardly established only by handiwork of ontology, because texts specific to the domain do not include any kind of noun phrase for labeling these new terms. The ninth chapter brings out problems of automated indexing methods and information fetching. Theoretically, it shows the LSA method and its application on the example of encyclopedical articles of Croatian Encyclopedia with the goal of learning about its effieciency and goals in building ontology of an certain area. Conducting LSA method on chosen articles shows its utility in assessing the term as a gathering of related terms and affiliations of certain entries of articles (documents) in that matter, by which on the basis of word forms that selected articles consist of allows an automatic classification of articles by individual ontologic classes. The tenth, also final, chapter of this work is a conclusion which combines theoretical and practical part of the work by giving a short review of research results and showing the possibility of establishing interoperability and connecting Linked Data concept of Croatian encyclopedistics with other structural sources of encyclopedical knowledge on the web ( e.g. DBpedia, Freebase, etc.).

Actions (login required)

View Item