| dc.description.abstract | Language as the primary tool of human communication, reflects the cultural and
linguistic diversity in Indonesia. People often use non-standard language,
especially on social media, giving rise to variations in spelling, slang, and
abbreviations. Research aims to translate non-formal language into formal
Indonesian using a Semi-Supervised Translation approach. The study focuses on
the context of non-formal language on social media, particularly platforms X
previously known as Twitter, Facebook, Instagram, and the like. The use of slang
and abbreviations is a prominent characteristic of non-formal language. Therefore,
a model system is built using transformer architecture to translate non-formal
language into formal language. Testing is conducted by analyzing five test data,
evaluating accuracy, and examining the effectiveness of the translation model.
The results indicate that the Semi Supervised Translation model can be used to
build a translator system with an accuracy of up to 96%. However, some nonformal
words or sentences and abbreviations are still challenging to transform into
a formal format. In conclusion, this research suggests that the translation of nonformal
language into formal Indonesian can be achieved using Semi-Supervised
Translation. | en_US |
