METEOR
Metric for Evaluation of Translation with Explicit ORdering
METEOR (Metric for Evaluation of Translation with Explicit ORdering) is a metric for automatic evaluation of machine translation that calculates the similarity between a machine translation output and a reference translation using n-grams.
Meteor evaluates a translation by computing a score based on explicit word-to-word matches between the translation and a given reference translation.
METEOR: An Automatic Metric for MT Evaluation with Improved Correlation with Human Judgments
Apart from exact word matching, METEOR adds matching for synonyms and simple morphological variants of a word.
METEOR takes into account both the precision and recall while evaluating a match.
Variants
- METEOR-NEXT
Note: The list is incomplete.