LinkedHumanities - Linking and Populating the Digital Humanities
Zusammenfassung der Projektergebnisse
In the first half of the project, we developed the linked open data enhancer (LODE) which is an easy-to-use web framework which allows non-experts to link their local repository to other sources and to enhance their data with information from these linked sources. Furthermore, we developed an intuitive search interface for linked data where users can easily search for data in the local repository and can easily filter their search results according to intuitive filter criteria. In the second half of the project, we further improved our searching, linking and enhancing interface turning the alpha version into a stable system. In particular, we generalized our framework such that any data in RDF format can be loaded. Besides InPhO data, we loaded data from the library of congress and the Never Ending Language Learning (Nell) project. The browser interface has been extended with additional functionality like auto-completion and clustering of individuals connected via same-as relations. Furthermore, LODE now allows to add many more semantic relations like disjoint with, broader than, related to, and many more. It incorporates two linking algorithms which work ”out of the box“ and which are well suited for suggesting high quality linking candidates. This has been shown in our experiments. In future work, we will perform a user study to evaluate the entire framework including the enhancement component. One aspect we aim to examine are comparisons between different (more sophisticated) context selection algorithms. Furthermore, we are continuously improving the linking and enhancement algorithms. Additionally, we plan to extend the enhancement interface so as to also allow the manual addition of novel content. We are also continuing advertizing LODE as valuable tool for other Linked Humanities projects to link and enrich their locally created RDF datasets with the LOD cloud. The focus will be on submitting tutorials/workshops to advertise the tools and help humanities scholars to use them. The demo and the full paper submitted help to increase the awareness of LODE within the Linked Humanities community and will encourage other projects to use LODE.
