Detecting Artificial Inflation of SMS Traffic Using Machine Learning

The rapid expansion of SMS traffic has led to an increase in artificial inflation, where entities artificially manipulate message volumes for various purposes. Detecting such inflation is essential for maintaining the integrity of communication systems, yet effective solutions to this problem remain limited. This paper introduces a machine learning approach to identify artificial inflation in SMS traffic. Several models were developed and evaluated-including Random Forest, Gradient Boosting, Support Vector Machine (SVM), and Gaussian Naive Bayes (GaussianNB)-using a dataset provided by a telecommunications company. The models were trained on features related to recipient numbers and other contextual factors, and their performance was compared on the basis of accuracy metrics. The results indicate that the Gradient Boosting model significantly outperforms the others, achieving an accuracy score of 91%, and shows considerable potential to effectively detect artificial