A Comprehensive Multimodal E-mail Spam Detection Farmwork— An analysis of significance and challenges

DOI

10.2991/978-94-6239-616-6_125How to use a DOI?

Keywords

Email spam detection; hybrid feature selection; multimodal learning; Transformer; CNN; federated learning; XAI; adversarial robustness; phishing; cybersecurity

Abstract

Email spam endures to pose a momentous cybersecurity threat by manipulating polymorphism, complication, and multimodal content to detour traditional detection mechanisms. Prevailing rule-based and single-modality-based machine learning methods characterize the limitations in becoming accustomed to budding spam campaigns, in handling high feature dimensionality, and in lower levels of interpretability. In this work, an experimental data set with 20 newsgroups has been generated and authenticated by high- performance metrics across ML algorithms like Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR), Ensemble, and Deep Learning models, with ensemble and hybrid methods frequently setting benchmarks for accuracy, precision, recall, and F1 scores. As the next level, this study has proposed a novel multimodal hybrid framework that has resulted from the integration of advanced feature selection techniques with both machine learning and deep learning models. This framework considers and processes data consisting of a hybrid combination of textual, structural, and behavioral, as well as visual features, spreading over a hybrid feature selection pipeline that syndicates several novel ways of filtering, embedding, wrapping, and using attention-based and genetic optimization methods to boost the discriminative power and diminish the redundancy factor. Selected features are glued through transformer encoders, CNNs, and ensemble classification models. The framework uses Federated learning and adversarial training to strengthen robustness, whereas XAI methods helps in model interpretability. This work further discusses the significance and challenges involved in the development of this framework and existing research gaps have been analyzed.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - S. Sri Saye Lakshmi
AU  - K. Shantha Kumari
PY  - 2026
DA  - 2026/03/31
TI  - A Comprehensive Multimodal E-mail Spam Detection Farmwork— An analysis of significance and challenges
BT  - Proceedings of the International Conference on Artificial Intelligence and Secure Data Analytics (ICAISDA 2025)
PB  - Atlantis Press
SP  - 1770
EP  - 1788
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-616-6_125
DO  - 10.2991/978-94-6239-616-6_125
ID  - Lakshmi2026
ER  -