Intrusion Detection Based On Neural Network Ensemble

With the rapid development and wide application of computer network technologies, computer network security is becoming more and more important. As an important and active security mechanism, Intrusion Detection will reinforce the traditional system mechanism. The existing intrusion detection technologies have the deficiency of higher false positive rate, higher false negative rate, and poor real-time performance. Especially high detection accuracy is usually based on abundant or self-contained training data.High learning ability and rapidly answer-finding ability of neural network make neural network become a new solution to the urgent problem of intrusion detection system. The Intrusion Detection System based on neural network can detect intrusion rapidly and find many new attacks, but the detection accuracy and small samples study capacity should be better. Neural network ensemble technology took full advantage of neural networks. Further more, it concerns new methods, techniques of each domain and their combination. Each of the singular techniques could overcome its weaknesses by acquiring others' strong points. By combining them effectively this technology always performs better than that using only one among those techniques.Considering about the accuracy of individual classifier and diversity among individual classifiers, this paper integrates BP neural network and support vector machine based on the theory of neural network integrationFirstly, from the angle of the training samples, a method for neural network ensemble is proposed based on fuzzy c-means clustering. Using fuzzy c-means function, a distributed function is constructed and based on it; data are sampled from training samples. Then these data are used as training set of individual neural networks, many individual neural networks constitute neural network ensemble and the output of the ensemble uses majority-voting method. Theoretical analyses and experimental results show that this neural network ensemble method can improve the performance of pattern recognition and system testing.Secondly, from the perspective of classifier structure, a method for neural network ensemble is proposed based on based on the BP neural network and support vector machines according to the theory of neural network integration. The components which have more diversity are selected and ensembled. The intrusion detection system model are proposed based on the above and the simulation experiment results of experiment demonstrate that apply a hybrid of BP-SVM in intrusion detection system has better ability for feature selection, small samples study and intrusions detection.