FortifyGuard Cybersecurity Model: Fortifying Cybersecurity through an Enhanced Marine Predators Algorithm and Transformer-Based Approach for Advanced Malware Detection

The importance of malware detection makes it imperative that robust models are developed to correctly identify the threats.However, modern malware is often too intricate and numerous to be coped with by the aforementioned techniques.This paper suggests a new model of malware detection which is a combination of an improved marine predator algorithm for feature extraction and a transformer neural network for classification purposes.Enhanced MPA utilizes adaptive hunger and Particle swarm optimization-based hybridization to search out high dimensional feature space efficiently.The Transformer has self-attention mechanisms that are well suited to capturing intricate feature interdependencies and is subjected to a specific loss function which has regularization during training to avoid overfitting.When tested on KronoDroid standard dataset, the innovative FortifyGuard cybersecurity model achieves a detection level of 99.79% and a false positive rate of 0.05%.This combination of the advanced EMPA with the latest in machine learning certainly provides a very strong and efficient solution to the hard problem of malware definition within the scope of information security .