Polish to English Statistical Machine Translation

TL;DR

This paper investigates how different training configurations and data sources impact the performance of a Polish to English statistical machine translation system for spoken language, using multiple evaluation metrics.

Contribution

It presents an analysis of training settings and data preparation effects on translation quality for Polish-English SMT systems.

Findings

Data quality and preparation significantly influence translation accuracy.

Different corpora sources yield varying translation performance.

Evaluation metrics show consistent trends across experiments.

Abstract

This research explores the effects of various training settings on a Polish to English Statistical Machine Translation system for spoken language. Various elements of the TED, Europarl, and OPUS parallel text corpora were used as the basis for training of language models, for development, tuning and testing of the translation system. The BLEU, NIST, METEOR and TER metrics were used to evaluate the effects of the data preparations on the translation results.