Developing an Efficient Architecture to Increase the Accuracy of Automatic Detection of Suspicious Patterns of Intrusion in Cloud Computing using Neural Network

Intrusion detection systems today has considered as one of the main mechanisms for achieving network and computer network security. When the idea of using these systems has introduced, due to its high processing load, it is welcome by the military and commercial milieu. Nowadays, with the significant advancement in the design and production of specific hardware-specific circuits and the development of modern architectures in the design and production of software, it is possible to use this idea and technology for a wide range of computer systems. A noteworthy point is about the issue of providing network security, cost issues, energy consumption, reliability of the security system and the like. The purpose of this study is to provide a new method based on intrusion detection systems and its various architectures, with the aim of increasing the accuracy of intrusion detection in cloud computing. In this study, a NKDD99 dataset has used for teaching and classifying tests. It has shown that the results can have a better performance than any other work that has researched with the available features. In this study, we combine the methods of choosing linear correlation and interactive information. Of the 41 features in this database, there are ultimately 21 features and were used in stratified algorithms. Different classification algorithms, including decision tree, random forest, CART algorithm and neural network, were applied to the data. The results of these methods have shown that the best accuracy obtained in these methods is through the neural network method with 99.98%.

کلیدواژه ها:

Effective Architecture ، Accuracy Enhancement ، Pattern Recognition Intrusion in Cloud Computing ، Neural Network.