Sentiment Analysis of Cyber Attacks in Bank Syariah Indonesia Using SVM and Indobert Method
Year: 2025
Type: article
Abstract: Bank Syariah Indonesia (BSI) is one of the Islamic banking institutions that operates based on Islamic principles in accordance with Islamic law and has obtained an operational license from the Dewan Syariah Nasional (DSN). The advancement of information technology brings unique risks to the banking industry, including BSI. One example is the ransomware attack experienced by BSI from May 8 to 11, 2023, where 15 million customer data and 1.5 terabytes of internal data were stolen, leading to significant public concern and response across various media platforms. This has the potential to affect public trust in the Islamic banking industry, particularly BSI. This research aims to analyze public sentiment on Twitter regarding the attack to identify the majority sentiment formed, as well as to compare the performance of the SVM and IndoBERT models in classifying sentiments. Additionally, this study reveals the topics present in the negative sentiments based on the classifications of both models through topic modeling using Latent Dirichlet Allocation (LDA). The results indicate that the majority of sentiments are negative, while IndoBERT shows better performance compared to SVM, with an accuracy of 85% and an F1-Score of 82%. The topics present in the negative sentiments classified by SVM include issues related to fund security as well as transfers and withdrawals, whereas the topics present in the negative sentiments classified by IndoBERT are more related to problems with mobile banking and fund withdrawals.
Authors: Chandra Apriyadi, Styawati Styawati
Institution: Universitas Teknokrat Indonesia
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4410468408