Publication
SIGIR 2009
Conference paper

Cross language name matching

View publication

Abstract

Cross language information retrieval methods are used to determine which segments of Arabic language documents match name-based English queries. We investigate and contrast a word-based translation model with a character-based transliteration model in order to handle spelling variation and previously unseen names. We measure performance by making a novel use of the training data from the 2007 ACE Entity Translation task.

Date

Publication

SIGIR 2009

Authors

Topics

Share