IMPLEMENTASI BIDIRECTIONAL LONG SHORT TERM MEMORY (BiLSTM) UNTUK MENDETEKSI DEPRESI PADA AUGMENTASI TWEET BAHASA INDONESIA MENGGUNAKAN EASY DATA AUGMENTATION (EDA) DAN BACK TRANSLATION

MARSELINA, NYAYU CHIKA and Desiani, Anita and Irmeilyana, Irmeilyana (2023) IMPLEMENTASI BIDIRECTIONAL LONG SHORT TERM MEMORY (BiLSTM) UNTUK MENDETEKSI DEPRESI PADA AUGMENTASI TWEET BAHASA INDONESIA MENGGUNAKAN EASY DATA AUGMENTATION (EDA) DAN BACK TRANSLATION. Undergraduate thesis, Sriwijaya University.

[thumbnail of RAMA_44021_08011281924027.pdf] Text
RAMA_44021_08011281924027.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (1MB) | Request a copy
[thumbnail of RAMA_44021_08011281924027_TURNITIN.pdf] Text
RAMA_44021_08011281924027_TURNITIN.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (9MB) | Request a copy
[thumbnail of RAMA_44021_08011281924027_0011127702_0017057403_01_front_ref.pdf]
Preview
Text
RAMA_44021_08011281924027_0011127702_0017057403_01_front_ref.pdf - Accepted Version
Available under License Creative Commons Public Domain Dedication.

Download (838kB) | Preview
[thumbnail of RAMA_44021_08011281924027_0011127702_0017057403_02.pdf] Text
RAMA_44021_08011281924027_0011127702_0017057403_02.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (349kB) | Request a copy
[thumbnail of RAMA_44021_08011281924027_0011127702_0017057403_03.pdf] Text
RAMA_44021_08011281924027_0011127702_0017057403_03.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (175kB) | Request a copy
[thumbnail of RAMA_44021_08011281924027_0011127702_0017057403_04.pdf] Text
RAMA_44021_08011281924027_0011127702_0017057403_04.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (627kB) | Request a copy
[thumbnail of RAMA_44021_08011281924027_0011127702_0017057403_05.pdf] Text
RAMA_44021_08011281924027_0011127702_0017057403_05.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (28kB) | Request a copy
[thumbnail of RAMA_44021_08011281924027_0011127702_0017057403_06_ref.pdf] Text
RAMA_44021_08011281924027_0011127702_0017057403_06_ref.pdf - Bibliography
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (189kB) | Request a copy
[thumbnail of RAMA_44021_08011281924027_0011127702_0017057403_07_lamp.pdf] Text
RAMA_44021_08011281924027_0011127702_0017057403_07_lamp.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (111kB) | Request a copy

Abstract

Depression is a mental disorder that can interfere with daily activities. One way that can be done to detect depression on social media, especially Twitter, is by classifying the text in the tweets that are shared. In classification, a large amount of data is needed so that the results obtained are optimal, but sometimes the availability of data is limited. Augmentation is one way that can be used to increase the amount of data. Several augmentation techniques that can be used, namely Easy Data Augmentation (EDA) and back translation. Augmentation techniques help deep learning-based classification methods work more optimally. Bidirectional Long Short Term Memory (BiLSTM) is a deep learning algorithm that can be used for text classification. The advantage of BiLSTM is that it can identify a pattern in a sentence because each word is processed sequentially. In this study, augmentation of Indonesian tweets used EDA and back translation followed by classification using the BiLSTM architecture. The stages in this study are data collection, text preprocessing, data augmentation, tokenization, pad sequences, training, testing, evaluation, analysis and interpretation of results, and making conclusions. The results obtained from this study, namely accuracy, precision, recall, and f1-score were 98.68% each. These results indicate that the BiLSTM architecture is able to classify text in determining depressed and not depressed classes very well.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Back Translation, Bidirectional Long Short Term Memory, Depresi, Easy Data Augmentation, Klasifikasi Teks
Subjects: Q Science > QA Mathematics > QA8.9-QA10.3 Computer science. Artificial intelligence. Computational complexity. Data structures (Computer scienc. Mathematical Logic and Formal Languages
Divisions: 08-Faculty of Mathematics and Natural Science > 44201-Mathematics (S1)
Depositing User: Chika Marselina Nyayu
Date Deposited: 20 Feb 2023 01:58
Last Modified: 20 Feb 2023 01:58
URI: http://repository.unsri.ac.id/id/eprint/89692

Actions (login required)

View Item View Item