Evaluation of English–Slovak Neural and Statistical Machine Translation

This study is focused on the comparison of phrase-based statistical machine translation (SMT) systems and neural (NMT) using automatic metrics for quality evaluation language pair English Slovak. As approach predecessor translation, it was assumed that network would generate results with a better quality. An experiment performed residuals to compare scores accuracy (BLEU_n) those translation. The showed assumption regardless system used confirmed. There were statistically significant differences between SMT NMT in favor based all BLEU_n scores. achieved journalistic texts from into Slovak, if trained general texts, such as Google Translate, or specific ones, European Commission’s (EC’s) tool, which specific-domain.