云计算环境下光纤激光网络异常数据的高精度分类

High precision classification of optical laser network abnormal data in cloud computing environment

传统基于距离的分类方法只能获取球状簇,不能对云计算环境下光纤激光网络通信数据形成的不规则形状簇进行聚类,无法实现光纤激光网络异常数据的准确分类。因此,提出基于DBSCAN的异常数据高精度分类方法,其包括训练过程和检测过程。训练过程中光纤激光传感器节点采集数据,通过sink节点将数据反馈给中心基站,中心基站采用DBSCAN算法对数据进行训练,采集有价值的环境特征集并将其传递给sink节点。检测过程中Sink节点将特征集传递给激光传感器节点,激光传感器节点运算检测数据同环境特征集内的核心点的欧几里德距离,若该距离高于DBSCAN算法的训练半径,则说明检测数据是异常数据。实验结果说明,所提方法具有较高的分类效率、分类精度和较低的分类能耗。

Traditional classification method shape clusters of optical fiber laser network in cl based on distance can only oud computing environment obtain globular clusters, not the irregular which cannot realize accurate classification of network abnormal data. Therefore, abnormal data precision classification method based on DBSCAN is put forward, which includes training and testing processes. In training process, the optical laser sensor collects data first and feeds back the data to sink node center through the base station. Then adopts the DBSCAN algorithm to train data. Collect and pass the valuable environmental features to the sink node. In testing process, the Sink node will pass the features to laser sensor nodes and the laser sensor node will calculate and test the Euclid distance of core center in the environmental features. If the distance is higher than the training radius of DBSCAN algorithm, the testing data is abnormal data. The experimental results indicate that the proposed method has high classification efficiency, accuracy and lower energy consumption.