Detection and Prediction of Abnormal Users in Cloud Network

Objectives: Cloud computing refers to an internet based computing that allows sharing data and resources, which provides create, configure and customize the applications online. Methods: To overcome this issue, we proposed an analyzing behavior of user’s traffic. First, construct a whole feature set by collecting features from users and feature selection is helps to choose set of accurate features where necessary features are selected and predicted for an abnormal user is done by Naive Bayes classification. Findings: Support Vector Machine (SVM) produces high accuracy and over fitting of data is applicable. This new method improves the efficiency and detection rate of the system in the analysis of traffic behavior of abnormal users. Applications: Email communication, banking, e-commerce and social networks.