论文部分内容阅读
与传统电信网络和无线传感网络相比,光网络具有更快的数据传输速度和更大的数据规模,因此面临恶意数据攻击的风险随之增加。当前用于典型恶意数据的检测方法,无法应对大规模网络数据的系统检测,综合检测率水平较低,提出基于分布式BP神经网络与遗传算法相结合的恶意数据自动检测方法研究,依据BP神经网络算法原理,构建恶意数据特征检测模型;为提高样本数据学习和训练的效果和完整性,对大规模的数据集进行分割和风险评估;将分割后的小样本数据输入分布式BP神经网络模型进行自动检测,同时引入遗传算法求解最优网络权值,解决了BP算法收敛速度慢,易陷入局部最优的不足,最终实现对光网络中典型恶意数据的精确检测。仿真数据结果证明,提出的自动检测方法具有较高的检测率,同时也能将检测的误报率和漏报率控制在一个较低水平。
Compared with traditional telecom networks and wireless sensor networks, optical networks have faster data transfer rates and larger data sizes, so the risk of malicious data attacks increases. The current detection methods for typical malicious data can not deal with large-scale network data detection system, the overall detection rate is low, based on the combination of distributed BP neural network and genetic algorithm malicious data automatic detection method based on BP neural Network algorithm principle to construct the malicious data feature detection model; in order to improve the efficiency and completeness of the sample data learning and training, the large-scale data set is segmented and evaluated; the small sample data is input into the distributed BP neural network model Automatic detection, genetic algorithm is also introduced to solve the optimal network weights, which solves the shortcomings of the BP algorithm convergence speed is slow, easy to fall into the local optimum, and ultimately achieve the accurate detection of typical malicious data in the optical network. Simulation results show that the proposed automatic detection method has high detection rate, and can also control the false positive rate and false negative rate of detection to a lower level.