Word Alignment Based on Corpora Using MIM
Wu Hong · Mini-micro Systems · 2004
Proposed a method of word alignment which is based on sentence level aligned bilingual corpus. This work is based on the set form of sentence pairs. Using MIM(minimum intersection model) to align words. This model includes an algorithm based on inverted index table and set operations. Using HFDTable (high frequency disturbance word table) to improve recall. The results of experiments indicate that this method rivals the word alignment based on mutual information and bilingual lexicon.