Abstract: Aiming at the problem that the parameters of Least Squares Support Vector Machines（LS-SVM） are uncertain in time series prediction, this paper utilizes immune clonal memetic algorithm which adopts the advantage of global search and local search to optimize the parameters of LS-SVM. Simulation results of Lorenz time sequence prediction and building energy consumption prediction show that the prediction accuracy of this optimization method is higher than genetic algorithm and grid search algorithm, and the comparison shows that the optimized LS-SVM produces better results than Support Vector Machines（SVM） and BP neural network.

Key words: time series prediction, Least Squares Support Vector Machines（LS-SVM）, memetic algorithm, energy prediction

摘要： 针对最小二乘支持向量机（LS-SVM）在时间序列预测中的参数不确定问题，在训练阶段，使用结合了全局搜索和局部搜索的免疫文化基因算法来进行参数寻优。实验中通过对Lorenz时间序列和建筑能耗的两组预测实验，对比了免疫文化基因算法、遗传算法和网格搜索算法对LS-SVM参数的优化效果，证明了免疫文化基因算法的优化效果最好，且LS-SVM的预测精度比支持向量机（SVM）和BP网络预测都要高。

关键词: 时间序列预测, 最小二乘支持向量机, 文化基因算法, 能耗预测