DETECTION OF MALAKA REGIONAL LANGUAGES USING THE CONVOLUTIONAL NEURAL NETWORK (CNN) METHOD
JOURNAL OF SCIENCE AND APPLIED ENGINEERING · 26.10.2026 · 9 (2) · 80-85
DOI: 10.31328/jsae.v9i2.7692 ↗Kurz gefasst
Fachartikel in der Zeitschrift JOURNAL OF SCIENCE AND APPLIED ENGINEERING, online erschienen am 26.10.2026. Verfasst von Redeptus Feryanto Riu und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Indonesia possesses an extraordinary diversity of regional languages, including those spoken in the Malaka region of East Nusa Tenggara Province. This study aims to automatically detect and classify three Malaka regional languages — Tetun, Dawan, and Rote — using the Convolutional Neural Network (CNN) method. A balanced dataset consisting of 600 text samples (200 per language) was collected from various digital sources. Simple text preprocessing — namely punctuation removal and lowercase conversion — was applied, followed by model training under nine data-split ratios ranging from 10:90 to 90:10, in order to analyze the effect of training-data proportion on model performance. The CNN architecture comprised an embedding layer, a Conv1D layer with 128 filters and a kernel size of 5 (ReLU activation), a global max-pooling layer, and a fully connected layer with softmax output. The model was compiled using the Adam optimizer with a learning rate of 0.1 and trained for 10 epochs with a batch size of 8. The best performance was achieved at the 70:30 training-to-testing ratio, with precision, recall, F1-score, and accuracy each reaching 0.97. These results demonstrate that CNN is highly effective for regional language classification and that the proportion of training data significantly influences model performance.
Zitieren
APA
Riu, R. F., Putra, R. P., & Istiadi (2026). DETECTION OF MALAKA REGIONAL LANGUAGES USING THE CONVOLUTIONAL NEURAL NETWORK (CNN) METHOD. JOURNAL OF SCIENCE AND APPLIED ENGINEERING, 9(2), 80-85. https://doi.org/10.31328/jsae.v9i2.7692
Vancouver
Riu RF, Putra RP, Istiadi. DETECTION OF MALAKA REGIONAL LANGUAGES USING THE CONVOLUTIONAL NEURAL NETWORK (CNN) METHOD. JOURNAL OF SCIENCE AND APPLIED ENGINEERING. 2026;9(2):80-85. doi:10.31328/jsae.v9i2.7692
BibTeX
@article{Riu2026,
author = {Riu, Redeptus Feryanto and Putra, Rangga Pahlevi and Istiadi},
title = {DETECTION OF MALAKA REGIONAL LANGUAGES USING THE CONVOLUTIONAL NEURAL NETWORK (CNN) METHOD},
journal = {JOURNAL OF SCIENCE AND APPLIED ENGINEERING},
year = {2026},
volume = {9},
number = {2},
pages = {80--85},
doi = {10.31328/jsae.v9i2.7692}
}
EUVIMED