Building Ontologies from Collaborative Knowledge Bases to Search and Interpret Multilingual Corpora
Yegin Genc, Elizabeth Lennon, Winter Mason, Jeffrey V. Nickerson · 2013
Tools and techniques that automate the interpretation of multilingual corpora are useful on many fronts; scholars, as an example, could use such tools to more readily pinpoint relevant articles from journals in a wide variety of languages. This work describes techniques to build and characterize ontologies using collaborative knowledge bases, e.g., Wikipedia. These ontologies can then be used to search and classify texts. Originally developed for monolingual corpora, we extend the approach to multilingual texts and test the methods with Mandarin scientific abstracts. The presented techniques provide a novel and efficient mechanism to obtain contextually rich ontologies and measure document relevancy within multilingual corpora. 1