论文部分内容阅读
异常检测是光纤通信中需要解决的核心问题,为了准确检测出光纤通信中的网络流量异常行为,保证光纤网络的正常运行,提出一种基于分形理论的光纤网络流量异常检测技术。首先采用分形理论对光纤网络流量进行分析,获得网络的延迟时间向量,然后正交基神经网络建立网络流量异常检测模型,并将模型参数看作一种函数优化问题,最后采用教学优化算法对参数优化问题求解。实验结果表明,本文方法不仅提高了光纤网络流量异常检测的精度,而且对一些异常行为检测效果要明显优于当前其它光纤网络异常检测方法。
Anomaly detection is the core problem to be solved in optical fiber communication. In order to accurately detect abnormal network traffic in optical fiber communication and ensure the normal operation of optical network, an optical fiber network traffic anomaly detection technique based on fractal theory is proposed. Firstly, the fractal theory is used to analyze the traffic of optical fiber network to get the delay vector of the network. Then, the orthogonal neural network is used to establish the network traffic anomaly detection model, and the model parameters are considered as a function optimization problem. At last, Solve optimization problems. Experimental results show that this method not only improves the accuracy of optical network traffic anomaly detection, but also significantly outperforms other current optical network anomaly detection methods for some abnormal behavior detection.