摘要： 针对现有混沌支持向量机回归模型存在流量预测效率低下的问题，利用差分进化(DE)算法、遗传算法和粒子群优化算法确定模型的径向基核函数系数、惩罚系数、不敏感系数等参数，在此基础上建立改进的混沌支持向量机回归模型进行流量预测。实例表明，相比其他启发式算法，DE算法能以较高的效率搜索到混沌支持向量机回归模型的最优参数，并且该模型具有较高的预测精度。

关键词: 网络流量预测, 混沌支持向量机, 差分进化算法, 粒子群优化算法

Abstract: Aiming the problem of chaos Support Vector Machine(SVM) regression model has low efficiency traffic prediction, Differential Evolution(DE) algorithm, Genetic Algorithm(GA) and Particle Swarm Optimization(PSO) algorithm are applied for determining the model’s parameters which includes radial basis kernel function coefficient, punishment coefficient, insensitive coefficient, and builds the improved chaos SVM model to predict traffic. Example show that DE algorithm can significantly reduce the time for determining the model compared to another heuristic algorithms, and the model has higher prediction accuracy.

Key words: network traffic prediction, chaos Support Vector Machine(SVM), Differential Evolution(DE) algorithm, Particle Swarm Optimization (PSO) algorithm

中图分类号: 

• TP393