【摘要】 面对层出不穷的网络攻击现状，设计了基于机器学习的电力智能光纤网络自动监测和预警方法。该方法的电力光纤网络信息数据采集单元利用不同类型的检测探针采集电力光纤网络风险信息后，将其传输到电力光纤网络监测单元内；该单元对电力光纤网络风险信息实施约简后，建立风险监测指标体系，并使用主成分分析方法得到电力光纤网络风险监测结果；将电力光纤网络风险监测结果输入到高斯混合模型内，利用该模型划分当前电力网络风险等级，在将该风险等级和电力光纤网络风险监测结果，同时输入到机器学习算法的支持向量机模型内，经过该模型迭代输出当前电力光纤网络监测结果对应的风险等级。实验表明：该方法具备显著的光纤网络风险信息约简能力，以及较强的监测能力和预警能力，应用效果较佳。更多还原

【Abstract】 In the face of endless network attacks, the automatic monitoring and early warning method of power intelligent optical fiber network based on machine learning is studied. The power optical fiber network information data acquisition unit uses different types of detection probes to collect the power optical fiber network risk information, and then transmits it to the power optical fiber network monitoring unit; After the risk information of power optical fiber network is reduced, the risk monitoring index system is established, and the risk monitoring results of power optical fiber network are obtained by using the principal component analysis method; Input the risk monitoring results of power optical fiber network into the Gaussian mixture model, use the model to divide the current power network risk level, input the risk level and power optical fiber network risk monitoring results into the support vector machine model of machine learning algorithm at the same time, and output the risk level corresponding to the current power optical fiber network monitoring results through iteration of the model. The experiment shows that this method has remarkable ability of risk information reduction in optical fiber network, as well as strong monitoring ability and early warning ability, and its application effect is better.更多还原