<![CDATA[One of the negative impacts of current digital advances is the increasing number of SMS spam. Spam SMS poses a security risk to users because they can contain malicious links or requests for personal information that are used for malware, smishing, or fraud attacks. However, with the various protection measures available, not all spam SMS can be classified and prevented effectively. However, this problem can be minimized by creating an anti-spam SMS model which aims to classify SMS types. So this research aims to classify types of SMS that contain spam and spam by applying the Naïve Bayes algorithm. In this study, the dataset consisted of 5572 records consisting of 2 categories, namely spam and ham. This algorithm is able to show satisfactory performance in differentiating spam and spam messages because, according to the diversity of literature, the Naïve Bayes algorithm is suitable for use in English language datasets. The evaluation model displays good results with accuracy reaching 93.2%, precision 93.7%, recall 93.2%, and F1-score 91.6%. In addition, analysis in the research using the Receiver Operating Characteristic (ROC) curve shows an accuracy rate of 97.3%, indicating that the model has very good performance in classifying spam in SMS messages. However, there is still room for improvement through the use of new methods and larger and more diverse data sets. This research has an important involvement in working on communication security and user experience in using short message services.]]>
Copyrights © 2024
