Evaluating Machine Translations Using mNCD

This paper introduces mNCD, a method for automatic evaluation of machine translations. The measure is based on normalized compression distance (NCD), a general information theoretic measure of string similarity, and flexible word matching provided by stemming and synonyms. The mNCD measure outperforms NCD in system-level correlation to human judgments in English.

• Type: Journal
• Year: 2010
• Where: ACL
• Authors: Marcus Dobrinkat, Tero Tapiovaara, Jaakko Väyrynen, Kimmo Kettunen
• Keywords: ACL 2010, Computational Linguistics, Information Theoretic Measure, MNCD Measure, System-level Correlation