Study On Anomaly Detection Based On Hybrid Intelligent Systems Using Immune Algorithm

Anomaly detection, as an essential component in the information security assurance framework, settles the issues in which traditional methods such as access control, identification authentication and firewall couldn't handle. However, current intrusion detection systems lack effectiveness, adaptability and extensibility, and especially they become ineffective in the face of new kind of attacks. Aimed at these shortcomings, this thesis takes a data-mining view to bring forward hybrid intelligent systems based on the immune algorithm, and it is put into the research of anomaly detection.After introducing the basic knowledge about anomaly detection, this thesis presents a mixed anomaly detection scheme based on data mining. The thesis summarizes the researching actuality and the production of application, then analyses the standard theories of anomaly detection deeply. Moreover, the thesis synthesizes and compares every classical algorithm. Based on the theory analysis, the author proposes a new HIS: R-FC-CSNN according to Rough Set,Clustering theory,Fuzzy Logic,Immune Algorithm and Artificial Neural Network. Firstly, R-FC-CSNN uses the Rough Set to reduce the data.And then it clusters the data used the Clustering Theory. After that it uses different and improved ANN which is improved by Cloning Select Algorithm to train. Subsequently the data trained is fabricated by fuzzy power. Lastly, the data is trained by another improved ANN which is improved by Cloning Select Algorithm and the whole process of training is completed. By six classical dataset in UCI, the validity of the model is tested.At last, we describe in the process of building many subsets hierarchy classification models from data provided by KDD99. And the data detect the action of intrusion by the hybrid intelligent systems based on the immune algorithm. Finally, the result proves that the model which is proposed has better using effect.