Bilingual topic taxonomy generation based on bilingual documents clustering

Bilingual taxonomy is one of key components of multilingual Ontology. In this paper, affinity propagation clustering algorithm is used to cluster bilingual documents collection and generate bilingual topic taxonomy. Two bilingual topic taxonomy generation methods, i.e. bilingual documents clustering before or after text feature reconstruction, are described. Dataset in two domains are tested and result shows that: according to net similarity, the result of documents clustering after feature reconstruction is better than that before feature reconstruction.