Van Huyssteen, Gerhard B., and Suléne Pilon. 2009. “Rule-based conversion of closely-related languages: A Dutch-to-Afrikaans convertor.” Proceedings of the 2009 Conference of the Pattern Recognition Association of South Africa:23-28.
Van Huyssteen & Pilon 2009
English: Afrikaans, closely-related languages, Dutch, human language technology, machine translation, POS tagging, recycling
Afrikaans: Afrikaans, hersiklering, masjienvertaling, mensetaaltegnologie, naby verwante tale, Nederlands, woordsoortetikettering
English: For fast-tracking the development of resources for resource-scarce languages; one could transfer existing technologies from one language to another well-sourced; closely-related language. In this contribution; we describe the development and performance of a rule-based Dutch-to-Afrikaans convertor; with the aim to transform Dutch text so that it looks more like an Afrikaans text (even though it might not even be a good Dutch translation). The rules we used is based on systematic orthographic; morphosyntactic and lexical differences between the two languages. We report on an accuracy of 71% on word-level; after minor optimisation with regard to iteration of rules. In a small-scale evaluation on running text; we obtain a BLEU score of 0.2519. We conclude that such a rule-based approach to conversion of closely-related languages holds much promise; with potential application in technology transfer (or even machine translation) between such languages.
Afrikaans:
In: English
On: Afrikaans and Dutch