Fraudulent SMS identification based on enhanced BERT

Yanlong Shu; Hao Li

doi:10.1117/12.2653454

27 September 2022 Fraudulent SMS identification based on enhanced BERT

Yanlong Shu, Hao Li

Proceedings Volume 12346, 2nd International Conference on Information Technology and Intelligent Control (CITIC 2022); 123460S (2022) https://doi.org/10.1117/12.2653454
Event: 2nd International Conference on Information Technology and Intelligent Control (CITIC 2022), 2022, Kunming, China

Abstract

With the popularity of SMS services, telecommunication fraud incidents occur frequently, which seriously threaten people's life and property safety. The existing fraudulent SMS recognition techniques have the problems of feature selection relying on manual and low recognition. To solve this problem, a fraudulent SMS identification method (referred to as BERT-SIG) that combines BERT (bidirectional encoder representation from transformers) and Sigmoid is designed. The method performs automatic feature extraction by multi-headed attention mechanism, improves the quality of recognition by Sigmoid activation function, and performs the task of recognizing fraudulent SMS by fine-tuning to improve the efficiency of fraudulent SMS recognition. In the experiments, two open-source datasets, "Spam Email" and "E-Mail classification NLP", are used to check the persistence and robustness, and the BERT-SIG is compared with various machine learning models. The experimental results show that BERT-SIG has the best performance accuracy and F1 score, reaching 99.72% and 99.53% respectively, and can effectively identify fraudulent SMS.

Citation Download Citation

Yanlong Shu and Hao Li "Fraudulent SMS identification based on enhanced BERT", Proc. SPIE 12346, 2nd International Conference on Information Technology and Intelligent Control (CITIC 2022), 123460S (27 September 2022); https://doi.org/10.1117/12.2653454

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available