Sentiment Analysis of Consumer Reviews on the Material Quality of Erigo Fashion Products on Shopee using IndoBERT
Year: 2026
Type: article
Abstract: Erigo is a local fashion brand that is widely sold on the Shopee marketplace, yet consumer opinions about its material quality are scattered across a large volume of unstructured reviews. This study aims to analyze consumer sentiment toward the material quality of Erigo fashion products based on Shopee reviews using the IndoBERT model, with a focus on three aspects: comfort, softness, and size fit. Review data were collected through web scraping and then processed through data selection, cleaning text, case folding, tokenization, sentiment labeling, aspect identification, and fine-tuning of the IndoBERT model to classify sentiment into positive, neutral, and negative. Of the 9,468 reviews analyzed, 4,967 were positive, 987 were neutral, and 3,514 were negative, so consumer sentiment toward the material quality of Erigo products is dominated by positive opinions. The model achieved an accuracy of 90.89%, a macro F1-score of 90.87%, and a ROC AUC of 96.87. The analysis results were implemented in a web-based information system developed with the Prototype method and verified with Black Box Testing, which presents structured information about consumer perception to support product quality evaluation and marketing strategy.
Authors: Amanah, Adhy Rizaldy, Syahbudin
Institution: Universitas Islam Negeri Alauddin Makassar
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7224849898