云计算环境下的光网络异常特征提取方法

A method for extracting the abnormal characteristics of optical networks in the cloud computing environment

基于异构环境的特征提取方法在特征分类过程中,不能解决云计算环境下光网络数据分类问题,导致光网络异常特征提取准确度较低。为获取可靠的光网络异常特征提取结果,分析云计算环境下的光网络异常特征提取方法,通过DBSCAN算法的训练过程获取光网络云计算环境特征集,在此基础上采用该算法的检测过程,完成云计算环境下云光网络异常数据分类;依据分类结果,采用碰撞思维获取光网络异常数据的差异性特征,以此为基础构建光网络异常数据的观测向量特征空间,在该向量特征空间内提取光网络异常特征。实验结果表明,所提方法提取光网络异常特征的精度均值高达94%,是一种有效的光网络异常特征提取方法。

The feature extraction method based on heterogeneous environment can not solve the problem of optical network data classification in the process of feature classification,which leads to the low accuracy of optical network anomaly feature extraction.In order to obtain reliable optical network anomaly feature extraction results,the optical network anomaly feature extraction method under cloud computing environment is analyzed,and the characteristic set of optical network cloud computing environment is obtained through the training process of DBSCAN algorithm.On this basis,the detection process of this algorithm is used.Complete the classification of Yunguang network anomaly data in cloud computing environment;According to the classification results,the difference characteristics of optical network anomaly data are obtained by using collision thinking.Based on this,the observation vector feature space of optical network anomaly data is constructed,and the optical network anomaly feature is extracted from the vector feature space.The experimental results show that the proposed method is an effective method to extract abnormal features of optical networks with an accuracy mean of up to 94%.