CAJ | 학술논문

为提高网络流量预测模型对稳定性和精度要求，提出了一种融合经验模式分解（EMD）和全局版人工鱼群LS-SVM模型的网络流量预测算法。该模型首先通过EMD分解，将原始流量的非平稳序列分解为若干不同尺度的平稳IMF（固有模态分量）分量，再用LS-SVM模型分别对各分量进行预测，其中LS-SVM的参数用全局版人工鱼群算法进行优化；最后通过SVM组合得到原始序列的预测值。仿真实验验证了该算法的预测准确性和稳定性。
위제고망락류량예측모형대은정성화정도요구，제출료일충융합경험모식분해（EMD）화전국판인공어군LS-SVM모형적망락류량예측산법。해모형수선통과EMD분해，장원시류량적비평은서렬분해위약간불동척도적평은IMF（고유모태분량）분량，재용LS-SVM모형분별대각분량진행예측，기중LS-SVM적삼수용전국판인공어군산법진행우화；최후통과SVM조합득도원시서렬적예측치。방진실험험증료해산법적예측준학성화은정성。
In order to improve the requirements of stability and accuracy for network traffic prediction model,a model that combines EMD and global artificial fish swarm algorithm of LS-SVM is proposed. Firstly the model is decomposed by EMD, making the original non-stationary flow sequence decomposing into different scales smooth IMF component, then using particle swarm optimization of LS-SVM to predict each component separately, and final-ly the predicted results of the original sequence through the SVM combination is acquired. The simulation results show that the method not only obtain higher prediction accuracy, but also reflect the mutability of original flow data.