TY  - THES
N1  - Ir. Maria Ulfah Siregar, S. Kom., MIT., Ph.D.
ID  - digilib78778
UR  - https://digilib.uin-suka.ac.id/id/eprint/78778/
A1  - Muhimmatul Mahmuda, NIM.: 24206051014
Y1  - 2026/08/13/
N2  - This study evaluates sentiment classification performance and detects echo chamber and filter bubble patterns in Indonesian marketplace product reviews. The research is motivated by the abundance of product reviews as valuable information assets whose collective implications on user information isolation remain largely unexplored. Using a dataset of 76,256 beauty product reviews, this study aims to compare the effectiveness of three classification models and map opinion dynamics within the digital ecosystem. The approach integrates sentiment analysis theory with social network analysis using Support Vector Machine (SVM), Logistic Regression, and IndoBERT. Information isolation patterns are detected using the Louvain Community Detection algorithm and six Vocabulary diversity metrics. Experimental results show that IndoBERT is the most optimal model with an accuracy of 81.17% and an F1-Score of 0.8123. Network analysis reveals moderate echo chamber and filter bubble phenomena with a Modularity Score of 0.2978 and the formation of 8 significant communities. These findings indicate early stages of opinion fragmentation where users tend to interact within homogeneous sentiment clusters but are not yet fully isolated. This research contributes to expanding the literature on information isolation in the Indonesian e-commerce domain and provides a practical framework for platform managers to monitor the health of user review interactions.
PB  - UIN SUNAN KALIJAGA YOGYAKARTA
KW  - Analisis Sentimen
KW  -  Echo chamber
KW  -  Filter bubble
KW  -  Marketplace
KW  -  IndoBERT
KW  -  Jaringan Kemiripan Teks
M1  - masters
TI  - EVALUASI HASIL ANALISIS SENTIMENT SERTA DETEKSI POLA ECHO CHAMBER DAN FILTER BUBBLE PADA ULASAN PRODUK MARKETPLACE
AV  - restricted
EP  - 142
ER  -