Semi-automatic Approach to Building Dictionary between Slavonic Languages

Machine translation between Slavonic languages is still in its early stages. Existence of bilingual dictionaries have big impact on quality of translation. Unfortunately creating such language resources is quite expensive. For small languages like Czech, Slovak or Slovenian is almost sure that large-enough dictionary will not be commercially successful. Slavonic languages tends to range between close and very close languages so it is possible to infer some translation pairs. Our presentation focus on describing semi-automatic approach using ‘cheap’ resources for CzechSlovak and Serbian-Slovenian dictionary. These resources are stacked so in earlier phases we will receive results of higher precision. Our results show that this approach improves effectivity of building dictionaries for close languages. Petr Sojka, Aleš Horák (Eds.): Proceedings of Recent Advances in Slavonic Natural Language Processing, RASLAN 2008, pp. 10–10, 2008. c ©Masaryk University, Brno 2008