摘要
提出一种应用人工神经网络进行入侵检测分类器设计的新方法,即采用改进的BP神经网络Levenberg-Marquardt优化算法进行入侵检测分类器的设计。该网络μ参量可自适应调整,收敛速度快,解决了传统BP算法的收敛速度慢、易陷入局部最小点的问题。实验结果表明,分类器用于入侵检测,效果良好,学习速度快,分类准确率高,为实现入侵检测分类器提供了一条准确高效的途径。
A new method of designing the classifier for intrusion detection is proposed based on neural networks,and the improved BP neural network is used which adopts LevenbergMarquardt(LM) optimized algorithm. In the algorithm, the parameter μ is selfadaptable, and the network convergence speed is high,and the problems of the conventional BP algorithm such as falling into the local minimum point easily and converging slowly are solved. The experimental result shows that the performance of the classifier for intrusion detection is favorable, the learning speed of the classifier is fast, and the rate of accurate classification is high,so it is practicable to design the classifier based on the BP neural network which adopts the LM algorithm. The network provides a precise and efficient way for implementing the classifier in intrusion detection.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
2003年第2期281-285,共5页
Journal of Hefei University of Technology：Natural Science
