Yanuar, Maria Ulfa (2025) Pemodelan Topik dan Analisis Sentimen pada Ulasan Pengguna Aplikasi Trans Jatim Menggunakan Latent Dirichlet Allocation dan Model IndoBERT. Masters thesis, Institut Teknologi Sepuluh Nopember.
|
Text
6032241147-Master_Thesis.pdf Restricted to Repository staff only Download (5MB) | Request a copy |
Abstract
Transportasi publik memegang peranan penting dalam mendukung mobilitas masyarakat, mengurangi kemacetan, dan menekan dampak lingkungan akibat penggunaan kendaraan pribadi. Untuk menunjang efektivitas layanan tersebut, Pemerintah Provinsi Jawa Timur menghadirkan aplikasi Trans Jatim Ajaib sebagai sarana digital yang memberikan informasi jadwal, rute, dan pelacakan bus Trans Jatim secara real-time. Keberhasilan aplikasi ini tidak hanya ditentukan oleh fitur teknologi, tetapi juga oleh pengalaman dan persepsi pengguna yang tercermin dalam ulasan di platform digital. Penelitian ini bertujuan untuk memodelkan topik ulasan pengguna aplikasi Trans Jatim menggunakan Latent Dirichlet Allocation (LDA) dan menganalisis sentimennya menggunakan model IndoBERT. Hasil analisis menunjukkan bahwa LDA berhasil mengidentifikasi 5 topik utama dengan fokus terbesar pada aspek kenyamanan dan jadwal rute. Analisis sentimen mengindikasikan mayoritas ulasan pengguna bersentimen positif, dengan aspek informasi aplikasi dan jalur memperoleh persepsi paling baik sedangkan aspek metode pembayaran menjadi area yang perlu diperbaiki. Model IndoBERT memiliki kinerja yang andal dalam mengklasifikasikan sentimen ulasan pengguna aplikasi Trans Jatim dengan akurasi sebesar 88,2%.
=================================================================================================================================
Public transportation plays an important role in supporting public mobility, reducing traffic congestion, and minimizing environmental impacts caused by private vehicles. To maintain the service effectiveness, the Government of East Java Province has presented Trans Jatim Ajaib application as a digital platform which provides real-time information concerning the schedules, routes, and Trans Jatim bus tracking. The success of this application is not only determined by its technological features, but also the user experience and perception as reflected in the reviews on the digital platform. This study aims to model user review topics of Trans Jatim application using Latent Dirichlet Allocation (LDA) and to analyze the sentiments using IndoBERT model. The analysis results demonstrated that LDA successfully identified 5 main topics with the primary focus on the aspect of convenience and route schedules. The sentiment analysis indicates that most of the user reviews carry a positive sentiment, with the aspect of information on the application and routes receiving the most favorable perception, while the payment method aspect becomes an area for improvement. The IndoBERT model showed reliable performance in classifying the user review sentiments on the Trans Jatim application with 88.2% accuracy.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Analisis Sentimen, Latent Dirichlet Allocation, Model IndoBERT, Pemodelan Topik, Trans Jatim, IndoBERT Model, Latent Dirichlet Allocation, Sentiment Analysis, Topic Modeling, Trans Jatim |
| Subjects: | T Technology > T Technology (General) > T57.5 Data Processing |
| Divisions: | Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT) |
| Depositing User: | Maria Ulfa Yanuar |
| Date Deposited: | 14 Jan 2026 07:52 |
| Last Modified: | 14 Jan 2026 07:52 |
| URI: | http://repository.its.ac.id/id/eprint/129624 |
Actions (login required)
![]() |
View Item |
