Iftinan, Jihan Hasna (2026) KOMPARASI METODE CLUSTERING PADA PIPELINE BERTOPIC DALAM ANALISIS SENTIMEN MULTI-ASPEK PADA APLIKASI MYSILOAM. Undergraduate thesis, UPN Veteran Jawa Timur.
|
Text (cover)
organized.pdf Download (982kB) | Preview |
|
|
Text (bab 1)
Bab 1.pdf Download (569kB) | Preview |
|
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (bab 2)
Bab 2.pdf Restricted to Repository staff only until 25 May 2029. Download (620kB) |
|
|
Text (bab 3)
Bab 3.pdf Restricted to Repository staff only until 25 May 2029. Download (1MB) |
|
|
Text (bab 4)
Bab 4.pdf Restricted to Repository staff only until 25 May 2029. Download (7MB) |
|
|
Text (bab 5)
Bab 5.pdf Download (301kB) | Preview |
|
|
Text (daftar pustaka)
Daftar Pustaka.pdf Download (306kB) | Preview |
|
|
Text (lampiran)
Lampiran.pdf Restricted to Repository staff only until 25 May 2029. Download (262kB) |
Abstract
The high volume of user reviews on digital health applications has not been optimally utilized to identify specific aspects affecting user experience. This study analyzes multi-aspect sentiment on MySiloam application reviews using a combination of BERTopic and Support Vector Machine (SVM). A total of 2,657 reviews were collected from Google Play Store and App Store spanning 2019–2025, filtered to 1,699 reviews after preprocessing. Aspect extraction was performed using BERTopic with a comparison of three clustering algorithms (HDBSCAN, BIRCH, K-Means), evaluated using topic modeling metrics namely C_v, UMass, NPMI, and Topic Diversity to assess topic coherence and diversity. Sentiment classification using SVM One-vs-One (OvO) was compared across two scenarios: a two-stage approach and a joint classification approach. K-Means with stemming produced the best topic quality (C_v=0.4113) and identified three service aspects. Medical Application Features, Service & Satisfaction, and Technical Application, with a labeling consistency of Krippendorff's Alpha 0.8816. The two-stage approach achieved the highest F1-score of 89.53% compared to the joint approach at 82.74%, demonstrating the superiority of task decomposition in sentiment classification. The combination of BERTopic K-Means with stemming and SVM OvO proved effective for aspect-based sentiment analysis on Indonesian digital health application reviews.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | Q Science > Q Science (General) Q Science > QA Mathematics Q Science > QA Mathematics > QA76.6 Computer Programming T Technology > T Technology (General) T Technology > T Technology (General) > T58.6-58.62 Management Information Systems |
||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Information Systems | ||||||||||||
| Depositing User: | Jihan Hasna Iftinan | ||||||||||||
| Date Deposited: | 26 May 2026 01:12 | ||||||||||||
| Last Modified: | 26 May 2026 02:13 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/52534 |
Actions (login required)
 |
View Item |