Detection Of Guilt Agent for Data Leakage In Cloud Environments

Abstract

Now-a-days many people are using internet widely and sharing of data through net has been increasing day by day. So there is high demand for securing the data and to prevent data leakage from being leaked especially in organisations where the data is shared to users. Data distributers should take care of the data and they can find who is responsible for the leakage. There are many methods for finding the guilt agents. Cloud can be used to prevent the data from being leaked. We would like to propose an alternative methodology to implement in real world and it is different from traditional methods. Traditional methods like watermarking, fake object allocation are used for finding the guilt agents, but this methods will not work if the guilt agent knows the fake objects. So the other methods for finding the guilt agents is to be implemented. There are many methods in existence but all other methods are overwritten by different algorithms for getting the better results. This paper proposes a technique to detect any leakages of data has been occurred or not and who leaked the data by using algorithms present in machine learning and by using Saleshandy website. Based on the time spent by each person which we got from Saleshandy we have calculated the accuracy of the person whether he/she is guilt or not. We  also used k-means clustering to the data of persons ,and compared the results with the previous results. By comparing we can conclude that the person is guilt or not. However the accuracy of the result may increase up to 80 percent.