ASPECT-BASED SENTIMENT ANALYSIS PADA DATA MULTIMODAL TWITTER UNTUK BENCHMARK PERFORMA PRODUK STUDI KASUS SMARTPHONE

Simbolon, Tegar Oktavianto
Wayuni, Eka Dyar
Afandi, Mohamad Irwan

ILTEK : Jurnal Teknologi
Vol. 20 No. 02 (2025): ILTEK : Jurnal Teknologi

The growing public opinions about smartphones on social media have driven the need for an informative and comprehensive sentiment analysis system. This study aims to develop a smartphone benchmark dashboard based on Aspect-Based Sentiment Analysis (ABSA) using multimodal data. The data consists of tweet texts and image URLs listed in a CSV file. The process involves text preprocessing (case folding, tokenization, stopword removal, and stemming) and Optical Character Recognition (OCR) to extract text from images. The tweet texts and OCR results are then combined and classified using the Support Vector Machine (SVM) algorithm to predict sentiment for each product aspect, such as camera, performance, and others. The results show that the SVM model performs well in predicting neutral and negative sentiments, although the identification of positive sentiment still needs improvement. The developed dashboard assists consumers in comparing products and serves as a reference for producers in improving product quality and marketing strategies.