Analogical Translation of Medical Words in Different Languages

摘要

Term translation has become a recurring need in many domains. This creates an interest for robust methods which can translate words in various languages. We propose a novel, analogy-based method to generate word translations. It relies on a partial bilingual lexicon and solves bilingual analogical equations to create candidate translations. We evaluate our approach on medical terms. To study the robustness of the method, we evaluate it on a series of datasets taken from different language groups and using different scripts. We investigate to which extend the approach can cope directly with multiword terms, and study its dependency to the size of the training set.