Performance of four sentence aligners on English Polish bitexts

Grzegorz Krynicki · Repozytorium Uniwersytetu im. Adama Mickiewicza (Adam Mickiewicz University in Poznań) · 2012

Collections of mutual translations gain in their value if aligned at the sentence level. As such they can be used e.g. in Statistical Machine Translation, Translation Studies and Lexicography. In this study, four leading systems of automatic sentence alignment are tested on four English-Polish translationally equivalent documents. Their performance is evaluated in terms of precision, recall and F-measure as well as in terms of their coverage of the source and target text. Some of the advantages and disadvantages of these methods with respect to different applications are discussed.

Read the paper · More papers on PaperTik