Naufal, Achmad (2025) Analisis Ulasan Pengguna Aplikasi Kesehatan Digital dengan Aspect-Based Sentiment Analysis sebagai Strategi Peningkatan Kualitas Layanan menggunakan Machine Learning. Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6032231182-Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6032231182-Master_Thesis.pdf - Submitted Version Restricted to Repository staff only until 1 April 2027. Download (7MB) | Request a copy |
Abstract
Perkembangan pesat layanan kesehatan digital telah secara signifikan mengubah cara pelayanan kesehatan diberikan dan dirasakan. Di Indonesia, platform seperti Halodoc berada di garis depan transformasi ini, menyediakan layanan telemedisin, konsultasi online, dan layanan kesehatan lainnya. Seiring dengan pertumbuhan platform ini, memahami dan meningkatkan pengalaman pengguna menjadi sangat penting. Ulasan pengguna dari Apple App Store memberikan informasi berharga tentang kepuasan pengguna serta area yang memerlukan perbaikan. Menganalisis umpan balik ini menjadi kunci untuk meningkatkan kualitas layanan dan beradaptasi dengan kebutuhan pengguna dalam ekosistem dinamis layanan kesehatan digital. Penelitian ini mengimplementasikan Aspect-Based Sentiment Analysis (ABSA) berbasis machine learning untuk menganalisis 4.793 ulasan pengguna di Apple App Store. Dengan menerapkan dan membandingkan berbagai model BERT, penelitian ini menghasilkan klasifikasi aspek dan sentimen yang efektif, dengan XLM-RoBERTa-Base mencapai Combined F1 score tertinggi (0,8686) dan IndoBERT-Base-Uncased menunjukkan performa kompetitif (0,8664) dengan efisiensi komputasi yang lebih baik. Analisis mencakup enam aspek utama layanan: kualitas layanan, kinerja aplikasi, pengalaman pengguna, pengiriman dan logistik, transaksi pembayaran, serta dukungan pelanggan. Tantangan seperti variabilitas bahasa dan kompleksitas umpan balik pengguna diatasi melalui kombinasi teknik Natural Language Processing (NLP) dan pendekatan hybrid dalam pelabelan data. Hasil penelitian mengungkapkan bahwa Service Quality mendominasi perhatian pengguna (47,6%), dengan distribusi sentimen menunjukkan 51,4% positif dan 46,2% negatif. Framework monitoring dan evaluasi yang dikembangkan memungkinkan transformasi dari pengelolaan informasi reaktif menjadi pendekatan berbasis data yang terstruktur. Penelitian ini menghasilkan rekomendasi konkret untuk peningkatan layanan, mencakup optimasi sistem pengiriman, stabilitas pembayaran, dan performa aplikasi. Metodologi dan temuan penelitian ini berkontribusi pada pengembangan sistem analisis umpan balik yang dapat diadaptasi untuk platform kesehatan digital lainnya, mendukung peningkatan kualitas layanan kesehatan digital di Indonesia.
================================================================================================================================
The rapid development of digital healthcare services has significantly transformed how healthcare is delivered and experienced. In Indonesia, platforms like Halodoc are at the forefront of this transformation, providing telemedicine services, online consultations, and other healthcare services. As these platforms grow, understanding and improving user experience becomes increasingly critical. User reviews from the Apple App Store provide valuable insights into user satisfaction and areas requiring improvement. Analyzing this feedback is key to enhancing service quality and adapting to user needs in the dynamic ecosystem of digital healthcare services. This research implemented machine learning-based Aspect-Based Sentiment Analysis (ABSA) to analyze 4,793 user reviews from the Apple App Store. By implementing and comparing various BERT models, the research achieved effective aspect and sentiment classification, with XLM-RoBERTa-Base achieving the highest Combined F1 score (0.8686) and IndoBERT-Base-Uncased showing competitive performance (0.8664) with better computational efficiency. The analysis covered six main service aspects: service quality, application performance, user experience, delivery and logistics, payment transactions, and customer support. Challenges such as language variability and complexity of user feedback were addressed through a combination of Natural Language Processing (NLP) techniques and a hybrid approach to data labeling. The research findings revealed that Service Quality dominated user attention (47.6%), with sentiment distribution showing 51.4% positive and 46.2% negative. The developed monitoring and evaluation framework enabled transformation from reactive information management to a structured data-driven approach. This research produced concrete recommendations for service improvements, encompassing delivery system optimization, payment stability, and application performance. The methodology and findings contribute to the development of feedback analysis systems that can be adapted for other digital healthcare platforms, supporting the improvement of digital healthcare services in Indonesia.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Aspect-Based Sentiment Analysis, BERT, Machine Learning, NLP, Halodoc |
| Subjects: | Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA336 Artificial Intelligence |
| Divisions: | Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT) |
| Depositing User: | Achmad Naufal |
| Date Deposited: | 31 Jan 2025 03:29 |
| Last Modified: | 31 Jan 2025 03:29 |
| URI: | http://repository.its.ac.id/id/eprint/117444 |
Actions (login required)
 |
View Item |