<![CDATA[This study investigates the classification of radical content in Indonesian Da’wah websites using text mining techniques. A content search engine application, developed with PHP, processes queries by comparing results against a database of keywords, classifying content into four categories: red, yellow, green, and white. Manual labeling based on data from the Ministry of Communication and Informatics yielded 126 labeled articles, forming the dataset for classification. The K-nearest neighbors (K-NN) algorithm, with an optimal k value of 7, achieved a classification accuracy of 66.37%, demonstrating its reliability compared to manual methods. The “White” class showed the highest precision and recall. System testing revealed efficient performance, with 0.704 seconds per classification task and 884,656 bytes of memory usage. Future enhancements include incorporating synonym identification for Indonesian keywords and exploring machine learning algorithms such as Naive Bayes and neural networks to improve accuracy. This research highlights the potential for text mining in identifying online radical content while emphasizing the need for system adaptability.]]>
