Abstract: In order to improve the predicting accuracy of network traffic, this paper proposes a network traffic predicting model（FCM-LSSVM） based on Least Square Support Vector Machine（LSSVM） and selecting train sample. Firstly, the Fuzzy C-Means（FCM） clustering algorithm is used to analyze the network traffic data to select the training samples and eliminate the isolated samples, and then the train samples are input to LSSVM to learn and the artificial bee colony algorithm is used to optimize the parameters of model, and finally, the predicting model is established and the simulation experiment is carried out to test the performance of the model. The simulation results show that the proposed has improved predicting accuracy of network traffic and speeding up modeling speed compared with other network traffic predicting models, so it can obtain the good predicting results.

Key words: network traffic, Least Squares Support Vector Machine（LSSVM）, Fuzzy C-Means（FCM） clustering, training samples

摘要： 为了提高网络流量的预测准确性，针对训练样本选取问题，提出一种训练样本选择的最小二乘支持向量机网络流量预测模型（FCM-LSSVM）。采用模糊均值聚类算法对网络充量数据进行了聚类分析，消除其中的孤立样本点，构建最小二乘支持向量机的训练集，然后将训练集输入到最小二乘支持向量机进行了学习，并采用人工蜂群算法对模型参数进行了优化，最后建立建立网络流量预测模型，并采用仿真实验对模型性能测试。仿真结果表明，相对于其他网络流量预测模型，FCM-LSSVM不仅提高了网络流量的预测精度，而且建模速度得以提高，获得了更加理想的网强流量预测结果。

关键词: 网络流量, 最小二乘支持向量机, 模糊均值聚类, 训练集