融合蚁群算法与支持向量机的网络流量预测

Prediction of network traffic by combination of ant colony optimization and support vector machine

网络流量的精确预测对控制网络拥塞有效控制有着重要意义。支持向量机是一种新的机器学习方法,能有效解决非线性、小样本及高维等问题。因为支持向量机的训练参数的取值与其预测能力有着较大关系,所以经常采用遗传算法选取训练参数。但是,遗传算法容易陷入局部极值,而蚁群算法具有全局优化能力。提出融合蚁群算法优化支持向量机,来提高网络流量预测精度。

Accurate prediction of network traffic is significant to control network congestion.Support vector machine is a new machine learning method,which can solve the problems with nonlinearity,small sample and higher dimension.Because the values of training parameters of support vector machine have close contact with its forecasting ability,so we usually choose training parameters using genetic algorithm.However,genetic algorithm is easy to fall into local extreme,and ant colony optimization has global optimization ability.Therefore,ant colony optimization and support vector machine（ACO-SVM） is proposed in the study to improve network traffic prediction accuracy.