Accessing multilingual data on the web for the semantic annotation of cultural heritage texts
Karlheinz Moerth, Thierry Declerck, Piroska Lendvai, Tamás Váradi · 2011
Abstract. Our study targets interoperable semantic annotation of Cultural Heritage or eHumanities texts in German and Hungarian. A semantic resource we focus on is the Thompson Motif-index of folk-literature (TMI), the labels of which are available only in English. We investigate the use lexical data on the Web in German and Hungarian for supporting semi-automatic translation of TMI: lexical resources offered by Wiktionary accessed via the Lexvo service, and discuss shortcomings of those resources. An approach for mapping the XML dump of Wiktionary onto a TEI and MAF compliant data is presented, whereby we discuss improvements in the representation of Wiktionary data for exploiting its multilingual value within the LOD framework.