Comparative Analysis of ML Techniques for Email Spam Detection
Juhi Singh, Prithvi Chaurasia, Yash Karnawat · 2025
Spam has increased as a result of the widespread use of email-based communication, making reliable and effective categorization techniques necessary to detect and filter undesirable information. Our paper presents a Semantic Graph Neural Network (SGNN) method that reframes email categorization as a graph-based problem. Instead of using conventional numerical embeddings, emails are represented as semantic graphs. This approach makes use of the relational and structural information included in email content, enabling a more sophisticated and contextually sensitive classification procedure. Our tests, which were carried out on a number of popular public datasets, show that SGNN routinely beats state-of the-art deep learning models and attains greater accuracy, especially in the difficult field of spam categorization. These findings highlight SGNN's promise as a practical and scalable approach to email spam detection in the real world, providing enhanced classification accuracy without the hassle of embedding layers.