Skripsi

PENGEMBANGAN MODEL PENERJEMAHAN BAHASA INGGRIS KE BAHASA JAWA MENGGUNAKAN LSTM DENGAN ATTENTION MECHANISM

---

One of the main challenges in English to Javanese machine translation is the model's difficulty in capturing contextual relationships between words, especially in complex sentences. Without an Attention Mechanism, the model tends to be unable to focus on the relevant parts of the source sentence during the translation process. This study aims to develop a Neural Machine Translation (NMT) model using the Long Short-Term Memory (LSTM) model based on Sequence-to-Sequence (Seq2Seq) architecture equipped with an Attention Mechanism to translate English sentences into Javanese. The research methodology involved training the model through six testing scenarios with varying hyperparameters, including LSTM units, batch size, learning rate, and dropout, to determine the most optimal configuration. Performance evaluation was conducted using the Bilingual Evaluation Understudy (BLEU) metric to measure lexical and structural accuracy. The results showed that Scenario 4, with a configuration of 128 LSTM Units, Learning Rate 0.005, and Dropout 0.5, produced the best and most stable performance compared to other models. This model achieved a BLEU-1 score of 72.63%, BLEU-2 of 53.67%, BLEU-3 of 41.05%, and BLEU-4 of 30.79%, indicating that the translation results fall into the understandable category with reasonably good grammatical structure. It is concluded that in limited datasets, adjusting the learning rate and regularization is more effective in improving translation accuracy and preventing overfitting than merely increasing the model architecture capacity.

---

Availability

Detail Information

Series Title

.

Call Number

T1960512026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xviii, 283 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.350 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Kecerdasan Buatan
Prodi Teknik Informatika

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment