CNN-BiLSTM Cloud Intrusion Detection Method Based on Contractive Auto-Encoder Feature Enhancement

The cloud computing environment is susceptible to various security threats, and ensuring security in cloud computing is imperative. Intrusion detection is an indispensable technology for protecting cloud environments. In this study, we employ a self- supervised feature-enhanced CNN-BiLSTM intrusion detection method to detect abnormal traffic in virtual flow data within a cloud environment. The approach utilizes information gain for feature selection, employs a shrinking auto encoder to enhance attack samples, and constructs a CAE-CNN-BiLSTM model to extract traffic features. These features are then input into a classification model for intrusion detection. Experimental results indicate that the proposed model achieves an Fl score and accuracy of 88.29% and 85.21 %, respectively, demonstrating its capability to enhance the precision of intrusion detection in cloud environments.