基于支持向量机的缓存预估模型的设计与实现

Design and implementation of buffer occupancy estimation model based on support vector machine

采用支持向量机(SVM)对网络业务流数据进行预测估计,通过训练样本,从而获得样本以外数据的分布规律。在此基础上,设计了一种网络排队队列缓存的预估模型。实验表明,该模型具有较高的训练效率和很高的估计精确度。

Based on the theory of support vector machine, distribution law of other data was obtained by analysis and training the sample data was obtained. On basis of this, a buffer estimation model was implemented. Simulation results show that this model have higher training efficiency and estimation precision.