Penerapan Lexicon-Based Untuk Analisis Sentiment Masyarakat Terhadap Kasus Kebocoran Data Di Indonesia

Abstract

The growth of social media has increased the availability of user-generated content that can be utilized for sentiment analysis, providing valuable insights into how the public perceives major social issues. One of the most critical issues widely discussed is cybercrime, particularly in the form of data leaks, which threaten individual privacy, institutional trust, and national security. Data leaks are increasingly recognized as part of a broader cybercrime landscape that demands public awareness as well as effective responses from stakeholders. Understanding how society reacts to this issue is therefore essential for government institutions and businesses. This study applies a lexicon-based method using three different sentiment dictionaries to classify sentiments expressed on Twitter regarding data leak issues. Data acquisition was conducted through web scraping, followed by preprocessing stages including case-folding, tokenization, stopword removal, and stemming. Sentiment classification was then performed to categorize tweets into positive, negative, and neutral classes. The results indicate that positive sentiment dominates with 2,866 tweets (45.07%), followed by negative sentiment with 2,298 tweets (36.12%), and neutral sentiment with 1,191 tweets (18.73%). These findings demonstrate that lexicon-based sentiment analysis, supported by multiple dictionaries, provides more reliable and comprehensive results than a single lexicon.