PARAPHRASE GENERATION UNTUK TEKS BAHASA INDONESIA MENGGUNAKAN LONG SHORT TERM MEMORY (LSTM)

RAFIKA, TARISA and Yusliani, Novi and Darmawahyuni, Annisa (2023) PARAPHRASE GENERATION UNTUK TEKS BAHASA INDONESIA MENGGUNAKAN LONG SHORT TERM MEMORY (LSTM). Undergraduate thesis, Sriwijaya University.

	Text RAMA_55201_09021281924076.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (19MB) | Request a copy
	Text RAMA_55201_09021281924076_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (4MB) | Request a copy
	Text RAMA_55201_09021281924076_0008118205_8968340022_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (4MB)
	Text RAMA_55201_09021281924076_0008118205_8968340022_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_55201_09021281924076_0008118205_8968340022_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_55201_09021281924076_0008118205_8968340022_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_55201_09021281924076_0008118205_8968340022_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (435kB) | Request a copy
	Text RAMA_55201_09021281924076_0008118205_8968340022_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (5MB) | Request a copy
	Text RAMA_55201_09021281924076_0008118205_8968340022_07_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (673kB) | Request a copy
	Text RAMA_55201_09021281924076_0008118205_8968340022_08_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (915kB) | Request a copy

Abstract

In the field of Natural Language Processing, there is a research area called Paraphrase Generation, which refers to the process of producing sentences that are semantically equivalent to the input sentence. With the advancement of neural methods, paraphrase generation, which previously relied on template-based approaches or static machine translation, can now utilize one of the neural methods, which is using Long Short Term Memory with the Sequence to Sequence model architecture, an architecture that consists of Encoder-Decoder layers. This study aims to investigate the performance of the LSTM model with the Sequence to Sequence architecture and incorporate the optimization technique using Attention in performing Paraphrase Generation for Indonesian language texts. The research results indicate that based on the evaluation of automatic metrics such as BLEU for each unigram, bigram, trigram, and quadgram, the model's scores are 0.48, 0.34, 0.23, and 0.15, respectively. Meanwhile, based on the evaluation of the automatic metric METEOR, the model's score is 0.51. In addition to the evaluation of automatic metrics, there is a questionnaire-based testing to assess the relevance and grammatical correctness of the paraphrased sentences generated by the model, based on human evaluation. The average scores for relevance and grammatical correctness, ranging from 1 to 5, are 3.71 and 4.10, respectively.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Paraphrase Generation, Long Short Term Memory, Sequence to Sequence, Attention, BLEU, METEOR
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics > Q325.5 Machine learning Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation. Q Science > Q Science (General) > Q350-390 Information theory
Divisions:	09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User:	Tarisa Rafika
Date Deposited:	03 Aug 2023 07:08
Last Modified:	03 Aug 2023 07:08
URI:	http://repository.unsri.ac.id/id/eprint/125426

Actions (login required)

View Item