An optimized and efficient approach for unknown attack detection using advance defence-GAN with PSO in cloud environment
Savita Devi, Taran Singh Bharti · 2024
In today&s;s increasingly interconnected society, cyber-attacks are a serious threat that must be addressed. Conventional machine learning has been applied to the problem of detecting network attacks in recent research. This was accomplished by studying the patterns of network behavior and then constructing a classification technique. These techniques often need enormous datasets that have been labelled; however, due to the unpredictability and rapid pace of cyber-attacks, it is impractical to label these datasets in real time. In order to solve these issues, a hybrid method called Advanced Defense GAN has been developed. This method uses Support Vector Classifier for Data Mining, Isolation Forest for Anomaly detection, and Particle Swarm Optimization for hyperparameter tuning. The goal of this method is to detect new and undiscovered attacks by using two datasets having known and unknown attacks. The accuracy score of the proposed hybrid technique is 99.96%.