The testing of this approach has used 25 single sentences without out-of-vocabulary (OOV), 25 single sentences with OOV, 25 plural sentences without OOV, and 25 plural sentences with OOV.
The next step is building NMT model and evaluate it. Then it continues to decide the NMT parameter model for the data training process. NMT Research has begun with creating a pair of 3000 parallel sentences of Lampung language (api dialect) and Indonesian language. The encoder in NMT is a recurrent neural network component that encrypts the source language to several length-stable vectors and the decoder is a recurrent neural networks component that generates translation result comprehensive. NMT, a new approach method in machine translation technology, that has worked by combining the encoder and decoder. In this research, automatically Lampung language translation into the Indonesian language was using neural machine translation (NMT) attention based approach.
0 kommentar(er)
