<![CDATA[This study presents a systematic review of sentiment analysis research conducted between 2021 and 2025 using the PRISMA method. From searches across three databases, namely Google Scholar, Semantic Scholar, and Garuda, a total of 12,089 articles were identified and then filtered down to 30 selected studies. The aim of this study is to identify the methods, algorithms, data sources, and accuracy levels used in sentiment analysis research. The findings indicate that the Naïve Bayes algorithm is the most widely applied, followed by SVM, while other algorithms such as KNN, Random Forest, Logistic Regression, and CNN were used only in limited cases. These findings highlight that sentiment analysis remains largely directed toward digital and social media issues, with classical algorithms such as Naïve Bayes and SVM continuing to be the main choices due to their ease of implementation and competitive accuracy.]]>
Copyrights © 2025
