Resources for Machine Translation of the Macedonian Language

Abstract

This paper focuses on creating new linguistic resources for the Macedonian language. It presents a new parallel corpus between Macedonian and Serbian language, build around the digitalized version of George Orwell's "1984", developed during the MULTEXT-EAST project. The original corpus is expanded with news articles from the Southeast European Times newspaper, published in public domain. The paper describes the retrieval, conversion, preprocessing, filtering and sentence-alignment of the corpus, then discusses and evaluates the alignment results.