A Study on the Mamba-ECANet Model for Intrusion Detection in Data Security Using End-to-End Learning

With the rapid advancement of information technology, network security has become an increasingly critical concern. In particular, data security intrusions pose significant risks to the privacy of data and the security of both enterprise and personal systems. Traditional intrusion detection systems often struggle with low detection accuracy and high false alarm rates, especially in complex and dynamic network environments with diverse attack techniques. To address these challenges, this paper proposes a deep learning-based data security intrusion detection system that integrates the Mamba model and the ECANet model, employing an end-to-end learning approach for training and optimization. First, the Mamba model is utilized for initial data feature extraction, offering efficient feature representation that lays a strong foundation for the detection process. Next, the ECANet model is incorporated to optimize feature selection using the attention mechanism, allowing the model to focus on the most critical features. Finally, the entire system is trained and optimized through an end-to-end learning approach, ensuring robust performance and reliability in real-world applications. Experimental results demonstrate that the proposed intrusion detection system achieves higher detection accuracy across various test datasets, with a 5% improvement over traditional methods, offering a novel and effective solution for data security.