Citation

Mohamad Ali, Nurulhuda and Goh, Hui Ngo and Lim, Amy Hui Lan and Man, Nurafiq Inani (2025) MPOS: Development of Malay Part-of-Speech Tagger using Embedding Methods and Neural Network Models. In: 2025 6th International Conference on Pattern Recognition and Machine Learning, PRML 2025, 13 June 2025 - 16 June 2025, Chongqing.

Text 14704.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

This paper presents the development of Part-of-Speech (POS) tagger for the Malaysian Malay language, employing the KDP-MPOS corpus as the foundational training dataset, whereby the corpus is the first that aligns with Kamus Dewan Perdana published by Dewan Bahasa dan Pustaka (DBP), a Malaysian government agency responsible for national Malay language development. To determine the best-performing combination, two types of word embeddings, Word2Vec and FastText, were integrated with four neural network models, RNN, LSTM, BiLSTM, and GRU. Experimental results demonstrate that the GRU model outperforms other models in performance for both word embeddings. It is observable that the combination of FastText with GRU resulted in the best performance by achieving accuracy 99.59% and F1 score 99.52%. While the combination of Word2Vec with GRU resulted in accuracy 99.44% and F1 Score 99.48%. Further analysis indicates that FastText performs better due to its ability to process the internal structure of words by breaking down each word into character n-grams, compared to Word2Vec’s approach of treating words as single entities, thus, highlight FastText's ability to better capture the morphological intricacies of the Malay language.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Deep learning, malay language, neural network model
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Computing and Informatics (FCI) Faculty of Applied Communication (FAC)
Depositing User:	Nurin Syazwani Azmi
Date Deposited:	06 Nov 2025 03:09
Last Modified:	06 Nov 2025 03:09
URII:	http://shdl.mmu.edu.my/id/eprint/14704

Downloads

Edit (login required)