부산대학교 도서관

Smishing is a type of cyberattack that includes delivering a fake messages to normal users in order to get their personal information. Cybersecurity experts are now concerned about this sort of assault because, both the individuals and companies are loosing their money by this attack. This article presents a machine learning base comparison approach using different text analysis features (i.e., Discourse, Linguistic, Psycholinguistic, Persuasion Principle and Statistics based features). Here, different Machine Learning classifiers (i.e., Logistic Regression, Naive Bayes, K-Nearest Neighbors, Random Forest, Support Vector Machine and XGBoost) have been used for smishing detection. The evaluation was performed on a prepared dataset. The dataset includes both smishing and normal messages. According to the comparative experimental results from the implemented classification algorithms, Random Forest classifier gives the best performance with the 99% accuracy.