Evaluation Metrics for Knowledge-Based Machine Translation

A methodology is presented for component-based machine translation (MT) evaluation through causal error analysis to complement existing global evaluation methods. Thismethodology is particularly appropriate for knowledge-based machine translation (KBMT) systems. After a discussion of MT evaluation criteria and the particular evaluation metrics proposed for KBMT, we apply this methodology to a large-scale application of the KANT machine translation system, and present some sample results.

• Type: Conference
• Year: 1994
• Where: COLING
• Authors: Eric Nyberg, Teruko Mitamura, Jaime G. Carbonell
• Keywords: COLING 1994, COLING 2008, Component-based Machine Translation, Evaluation, Machine Translation