SISTEM PENERJEMAH FRASA BAHASA BATAK TOBA BAHASA INDONESIA MENGGUNAKAN ALGORITMA RNN ENCODER -DECODER
Daftar Isi:
- Translation of the phrase Batak Toba - Indonesian is very important to be used as a reference system for the translation of the sentence Batak Toba - Indonesian. Recurrent Neural Network Encoder - Decoder algorithm is one of the techniques to build an automatic translator system. Translation is done by training the system to recognize the structure of phrases in the Batak Toba and Indonesian languages. So that formed a pattern that is the structure of both languages. Phrase translation is done by pre-processing, normalizing, and training with the Recurrent Neural Network encoder-decoder algorithm, and updating each weight generated with Backpropagation. The results of the experiments conducted obtained 0.8943 results using the Bilingual Evaluation Understanding (BLEU) Score. The data used in this study are 160 Batak Toba language phrases - Indonesian. With an accuracy of 89.43%. Keywords: Recurrent Neural Network Encoder - Decoder, Backpropagation, Bilingual Evaluation Understanding (BLUE) score