Leveraging LSTM Networks for SPAM Detection in Real-Time Chat Messages
Maitreyee Joshi, Mahendra Prabhakar Deore · 2025
SMS messaging is now more common than ever thanks to recent advancements in mobile technology.. As a result, there are now more spam communications are happening via mobile devices. Even though emails are the primary source of spam worldwide, SMS services are now not far behind in their contribution to the problem. No one wants their mobile devices to be inundated with spam messages. Numerous methods have been developed to identify and minimise spam messages, and research on the subject is constantly ongoing. Classifying spam in SMS messages presents significant challenges. Numerous studies have explored this problem through various machine learning techniques, such as Naive Bayes (NB), Random Forest (RF), and Support Vector Machine (SVM). These techniques, however, perform only to a limited extent, thus failing to classify a diverse variety of spam messages correctly. So, a thorough investigation is required to discover a more reliable and accurate way. To address this challenge, we propose a method known as Long Short-Term Memory (LSTM), an advanced architecture of Recurrent Neural Networks (RNN) that incorporates memory cells within its Gating Mechanism.