A Spam Transformer Model for SMS Spam Detection

In this paper, we aim to explore the possibility of Transformer model in detecting spam Short Message Service (SMS) messages by proposing a modified that is designed for SMS messages. The evaluation our proposed performed on Spam Collection v.1 dataset and UtkMl's Twitter Detection Competition dataset, with benchmark multiple established machine learning classifiers state-of-the-art detection approaches. comparison all other candidates, experiments show has optimal results accuracy, recall, F1-Score values 98.92%, 0.9451, 0.9613, respectively. Besides, also achieves good performance which indicates promising adapting similar problems.