Investigation of Machine Learning Approaches on Security Analysis of Cryptographic Algorithms

Abstract

The distributed denial of service (DDoS) assault was a kind of intrusion in the cloud computing environment that severely affects the end user by injecting illegitimate packets. To obtain performance, a hybrid improved wolf optimizer with asymmetric key Goldwasser cryptography (IWO-AKGC) algorithm was proposed based on combining the exploitation ability of security and exploration capability of machine learning. In addition to the selection of parameters, a proposed hybrid IWO-AKGC technique is used for weighting and bias coefficients in neural network models. This has led to an immediate improvement in communication security for the delivery of different types of data services via clouds, thanks to the proposed IWO-AKGC method. The recommended hybrid optimizer successfully addresses the drawbacks of conventional methods, such as local stagnation problems, delayed convergence problems, and local and global optimal trapping problems. Thus, secured data communication is obtained for cloud service provisioning. The proposed model proved to be a better model for DDoS intrusion detection.