Konstantinos Kyprianos

Highlighting Timely Information in Libraries Through Social and Semantic Web Technologies

Until now, timeliness of information in libraries is commonly used to underpin collection development and is directly related to quality in terms of realizing whether information is sufficiently up-to-date and available for use.

Providing LOD-Based Functionality in Digital Libraries

Libraries and other memory institutions have wasted no time in realizing that linked data technologies provide the necessary means to solve important interoperability issues that have been plaguing the community for decades. Despite the wide availability of cultural heritage information as linked open data – LOD, there seems to be a lack of LOD services that are targeted towards the end-user. In this paper, a LOD-powered, subject-based browsing service is proposed, capable of integrating resources from diverse repositories. More specifically, the proposed work describes a service that is built on top of a DSpace-based digital library of thesis and dissertations that not only exposes its topical information (i.e. subject headings) to the wider linked data community, but also manages to provide its end-users with additional relevant resources originating from a remote repository (i.e. New York Times – NYT articles database). The proposed service has been accordingly evaluated through a user survey.

Merging Controlled Vocabularies for More Efficient Subject-Based IR Systems

One of the most important tasks of a librarian is the assignment of appropriate subjects to a resource within a librarys collection. The subjects usually belong to a controlled vocabulary that is specifically designed for such a task. The most widely adopted controlled vocabulary across libraries around the world is the Library of Congress Subject Headings LCSH. However, there seems to be a shifting from traditional LCSH to modern thesauri. In this paper, a methodology is proposed, capable of incorporating thesauri into existing LCSH-based Information Retrieval-IR systems. In order to achieve this, a mapping methodology is proposed capable of providing a common structure consisting of terms belonging to LCSH and/or a thesaurus. The structure is modeled as a Simple Knowledge Organization System SKOS ontology, which can be employed by appropriate subject-based IR systems. As a proof of concept, the proposed methodology is applied to the DSpace-based University of Piraeus digital library.

Employing Twitter Hashtags and Linked Data to Suggest Trending Resources in a Digital Library

It is common truth that social web sites have dominated the web during the past few years. This results in the creation of vast amounts of information that is being produced by the corresponding user activities. Traditional information organization tools originating from the library domain are not applicable to the social web due to its overwhelmingly dynamic nature. Along these lines, hashtags have become an information organization tool of growing popularity among social web sites.

Merging Controlled Vocabularies through Semantic Alignment Based on Linked Data

In this paper, a methodology is presented that aids in finding equivalent terms between semantically similar controlled vocabularies. It is based both on lexical similarities discovery and semantic alignment through external LOD datasets. The proposed methodology has been deployed for the identification of equivalent terms within two datasets consisting of subject headings, namely Dione and NYT and facilitated through the employment of the LOD datasets of DBpedia and WordNet. The effectiveness of the methodology is assessed through a comparative evaluation between the deployment of the methodology presented in this paper and the deployment of a lexical similarities-based algorithm presented in previous work.