ANSWER GENERATION MENGGUNAKAN MODEL T5 BERDASARKAN DATA TRANSKRIP VIDEO YOUTUBE UNTUK PEMBELAJARAN DIGITAL

YUDHA, EKA WIRA and Abdiansah, Abdiansah (2025) ANSWER GENERATION MENGGUNAKAN MODEL T5 BERDASARKAN DATA TRANSKRIP VIDEO YOUTUBE UNTUK PEMBELAJARAN DIGITAL. Undergraduate thesis, Sriwijaya University.

	Image RAMA_55201_09021182126031_cover.pdf - Cover Image Available under License Creative Commons Public Domain Dedication. Download (164kB)
	Text RAMA_55201_09021182126031.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_55201_09021182126031_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (7MB) | Request a copy
	Text RAMA_55201_09021182126031_0001108401_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (3MB)
	Text RAMA_55201_09021182126031_0001108401_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (397kB) | Request a copy
	Text RAMA_55201_09021182126031_0001108401_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (447kB) | Request a copy
	Text RAMA_55201_09021182126031_0001108401_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (470kB) | Request a copy
	Text RAMA_55201_09021182126031_0001108401_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (740kB) | Request a copy
	Text RAMA_55201_09021182126031_0001108401_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (307kB) | Request a copy
	Text RAMA_55201_09021182126031_0001108401_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (181kB) | Request a copy

Abstract

Digital learning is rapidly evolving with the integration of technologies such as YouTube as a primary source of educational content. However, a major challenge remains in how learners can efficiently retrieve answers to specific questions from video transcripts. This study aims to design a question-answering system based on the T5 (Text-to-Text Transfer Transformer) model with input derived from YouTube video transcripts. The model is developed to extract relevant information and provide answers based on user queries. The methodology includes video transcript processing, and evaluating performance using F1-score, precision, and recall metrics. The results indicate that the best configuration was achieved with a batch size of 10, a learning rate of 1e-4, and a weight decay of 0.02, yielding an F1-score of 67.91%, recall of 88.64%, and precision of 68.07%. These findings demonstrate that the T5 model can generate fairly accurate answers based on YouTube video transcripts. The developed system is expected to serve as an effective solution in supporting more interactive and adaptive digital learning.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Answer Generation, T5, Pembelajaran Digital, Pemrosesan Bahasa Alami, YouTube.
Subjects:	B Philosophy. Psychology. Religion > BQ Buddhism > BQ135 Questions and answers. Maxims (General) L Education > L Education (General) > L7-991 Education (General)
Divisions:	09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User:	Eka Wira Yudha
Date Deposited:	12 Aug 2025 01:41
Last Modified:	12 Aug 2025 01:41
URI:	http://repository.unsri.ac.id/id/eprint/182554

Actions (login required)

View Item