Co-occurrence Graph Based Iterative Bilingual Lexicon Extraction From Comparable Corpora
Diptesh Chatterjee, Sudeshna Sarkar, Arpit Mishra · 2010
This paper presents an iterative algorithm for bilingual lexicon extraction from comparable corpora. It is based on a bagof-words model generated at the level of sentences. We present our results of experimentation on corpora of multiple degrees of comparability derived from the FIRE 2010 dataset. Evaluation results on 100 nouns shows that this method outperforms the standard context-vector based approaches.