IMPLEMENTASI ARSITEKTUR BIDIRECTIONAL LONG SHORT-TERM MEMORY DAN TRANSFORMER PADA KLASIFIKASI BERITA HOAKS DENGAN AUGMENTASI BACK TRANSLATION DAN TEXTATTACK

NURHALIZA, SITI and Desiani, Anita and Suprihatin, Bambang (2025) IMPLEMENTASI ARSITEKTUR BIDIRECTIONAL LONG SHORT-TERM MEMORY DAN TRANSFORMER PADA KLASIFIKASI BERITA HOAKS DENGAN AUGMENTASI BACK TRANSLATION DAN TEXTATTACK. Undergraduate thesis, Sriwijaya University.

	Text RAMA_44201_08011382126093.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_44201_08011382126093_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (10MB) | Request a copy
	Text RAMA_44201_08011382126093_0011127702_0026017102_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (929kB)
	Text RAMA_44201_08011382126093_0011127702_0026017102_02.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (569kB)
	Text RAMA_44201_08011382126093_0011127702_0026017102_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (236kB) | Request a copy
	Text RAMA_44201_08011382126093_0011127702_0026017102_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_44201_08011382126093_0011127702_0026017102_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (222kB) | Request a copy
	Text RAMA_44201_08011382126093_0011127702_0026017102_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (206kB) | Request a copy

Abstract

Hoax refers to false or misleading information. The spread of hoaxes can have a negative impact on society, such as disrupting stability and order. Automatic classification can help early detection of hoax. This study combines augmentation and classification methods. In the classification stage, this study proposes a combination of BiLSTM and Transformer architecture. BiLSTM is used to understand the order between words and Transformer is used to understand the global context. The combination of architectures requires large data. To fulfill the need for large data, this study proposes the implementation of augmentation using Back Translation and TextAttack. Back Translation and TextAttack are used to make the data augmentation results more varied. Performance evaluation is conducted by measuring accuracy, precision, recall, and F1-score. The application of the BiLSTM and Transformer combination to the augmented data achieved an accuracy of 98.80%, precision of 98.72%, recall of 98.87%, and F1-score of 98.79%. The precision and recall results with an average of 98% show that the proposed method is very good at predicting hoax and valid classes. The high F1-score value shows that the proposed method is not only able to provide excellent predictions, but also consistent in identifying hoax and valid classes. The use of the augmentation method increased the average by 3% for precision, recall, and f1-score, and increased accuracy by 0.56%. Based on these results, it can be concluded that the use of augmentation methods and the combination of BiLSTM and Transformer enables effective detection of hoax because it can distinguish news between hoax and valid classes.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Back Translation, Berita Hoaks, BiLSTM, TextAttact, Transformer
Subjects:	Q Science > QA Mathematics > QA75-76.95 Calculating machines > QA76.9.B45 Big data. Machine learning. Quantitative research. Metaheuristics.
Divisions:	08-Faculty of Mathematics and Natural Science > 44201-Mathematics (S1)
Depositing User:	Siti Nurhaliza
Date Deposited:	30 Jan 2025 08:18
Last Modified:	30 Jan 2025 08:18
URI:	http://repository.unsri.ac.id/id/eprint/167333

Actions (login required)

View Item