Dataset of evaluation error-rate metrics for journalistic texts EN/SK and DE/SK

Description

The dataset contains automatic metrics calculated for English and German journalistic texts translated into Slovak using four machine translation systems (Google Translate Statistical Machine Translation, Google Translate Neural Machine Translation, MT@EC Statistical Machine Translation, and E-Translation Neural Machine Translation). The automatic evaluation metrics are methods and techniques used to evaluate the quality of machine translation. This dataset contains scores of automatic metrics of error-rate that are from interval 0 to 1 where 0 represents a translation without errors and 1 a translation with many errors. The dataset contains values for each machine translation system.

Categories

Licence