Cloud Intrusion Detection Method Based on Stacked Contractive Auto-Encoder and Support Vector Machine

Security issues have resulted in severe damage to the cloud computing environment, adversely affecting healthy and sustainable development of computing. Intrusion detection is one technologies for protecting environment from malicious attacks. However, network traffic characterized by large scale, high dimensionality, redundancy, these characteristics pose serious challenges intrusion systems. Deep learning technology has shown considerable potential detection. Therefore, this study aims use deep extract essential feature representations automatically realize performance efficiently. An effective stacked contractive autoencoder (SCAE) method presented unsupervised extraction. By using SCAE method, better robust low-dimensional features can be learned raw traffic. A novel system designed on basis support vector machine (SVM) classification algorithm. The SCAE+SVM approach combines both shallow techniques, it fully exploits their advantages significantly reduce analytical overhead. Experiments show that proposed achieves higher compared three other state-of-the-art methods two well-known evaluation datasets, namely KDD Cup 99 NSL-KDD.