为了有效提升非结构化离散网络的传输质量,需要进行非结构化离散网络传输路径预测.但是采用当前方法进行非结构化离散网络传输路径预测时,忽略了网络传输信道和路径间的干扰影响,存在路径预测误差大的问题.为此,提出一种基于支持向量机回归的非结构化离散网络传输路径预测方法.方法先采用测量往返时延的基本思想进行非结构化离散网络传输路径时延测量,得到网络传输路径时延的抖动值,计算出网络传输路径的丢包率,获取发送端在网络传输路径上发送的最大数据量,在得到群数据包延迟序列后,对非结构化离散网络传输路径预测问题进行建模,结合支持向量机回归理论对模型进行求解,有效完成了对非结构化离散网络传输路径预测.仿真证明,所提方法能够有效提升非结构化离散网络传输路径预测精度,且预测效率较高.%In this paper,we propose a method for prediction of transmission path in unstructured discrete network based on regression of support vector machine (SVM).Firstly,delay of transmission path was measured with basic idea of measuring round-trip delay.Value of jitter of the delay was obtained.Then,packet loss probability of transmission path was worked out and maximum data volume transmitted by sending end on the path was acquired.Model of path prediction was built after obtaining the delayed sequence of clustered data package.Integrated with regression of SVM,the model was solved.Finally,the prediction was completed effectively.Simulation proves that the method can improve the prediction precision effectively,and the efficiency of prediction is higher.
展开▼
