Cloud Malicious Session Detection by Moth Flame Optimization Algorithm

Abstract

Computer network provide connection among various sector for different operation. Management of many information related work is done by cloud environment. Hence chance of attack on such virtual computers need to improve by development of security, monitoring software / algorithms. This paper has developed a cloud malicious session detection model by enhancing the training dataset using dimension reduction technique. Input dataset feature set were select in the basis of moth flame optimization algorithm. Due to dynamic adoption of moth flame algorithm for path selection out of many available proposed model use it for  dimension reduction. Selected features were used for training of multilayer neural network. Experiment was done on real dataset UNSW-NB15, this has cloud sessions with different attacks. Result shows tht proposed model has increased the comparing parameter values as compared to other existing models.