A concept-based inter-lingua and its applications to multilingual text and web mining

Abstract

In this paper, a novel concept-based approach for knowledge discovery from multilingual textual data is proposed and the potential applications of this approach in multilingual document mining and multilingual Web content mining are highlighted. The central idea is the generation of a concept-based inter-lingua encoding the multilingual concept-term relationship, using self-organizing map. Applying this concept-based inter-lingua as the multilingual linguistic knowledge base, conceptual content of arbitrary multilingual texts is then revealed using a novel multilingual conceptual indexing technique. © Springer-Verlag 2003.