Hybrid Fine Tuning IndoBERT dan LLM untuk Pelabelan Otomatis Aspect-Based Sentiment Analysis pada Review Aplikasi Mobile Banking

Natalia, Brenda (2026) Hybrid Fine Tuning IndoBERT dan LLM untuk Pelabelan Otomatis Aspect-Based Sentiment Analysis pada Review Aplikasi Mobile Banking. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6025242003-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (11MB) | Request a copy

Abstract

Ulasan pengguna terhadap aplikasi mobile banking yang tersedia di platform digital seperti Google Play Store menyimpan beragam informasi berharga yang dapat menjadi dasar peningkatan kualitas layanan. Namun, besarnya volume data dan terbatasnya ketersediaan dataset berlabel sering kali menjadi hambatan dalam proses analisis yang efisien. Sementara itu, pelabelan manual membutuhkan waktu dan tenaga yang tidak sedikit. Untuk menjawab tantangan tersebut, penelitian ini menawarkan pendekatan pelabelan otomatis berbasis Aspect-Based Sentiment Analysis (ABSA) dengan memanfaatkan model deep learning. Penelitian ini secara khusus membandingkan performa berbagai varian IndoBERT dan Large Language Models (LLM), serta mengevaluasi efektivitas model hybrid yang menggabungkan keduanya.
=================================================================================================================================
User reviews of mobile banking apps available on digital platforms like the Google Play Store contain a wealth of valuable information that can inform service quality improvements. However, the sheer volume of data and the limited availability of labeled datasets often hinder efficient analysis. Meanwhile, manual labeling is time-consuming and labor-intensive. To address these challenges, this study proposes an automated labeling approach based on Aspect-Based Sentiment Analysis (ABSA) utilizing deep learning models. Specifically, the study compares the performance of various variants of IndoBERT and Large Language Models (LLM) and evaluates the effectiveness of a hybrid model that combines both of them.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Pelabelan Otomatis, ABSA, IndoBERT, LLM, Mobile Banking, Auto-labeling
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55101-(S2) Master Thesis
Depositing User:	Brenda Natalia
Date Deposited:	01 Aug 2026 06:33
Last Modified:	01 Aug 2026 06:33
URI:	http://repository.its.ac.id/id/eprint/141046

Actions (login required)

View Item