支持向量机时间序列预测模型的参数影响分析与自适应优化

Influence analysis and self-adaptive optimization of support vector machine time series forecasting model parameters

建立在统计学习理论和结构风险最小原则上的支持向量机在理论上保证了模型的最大泛化能力,因此与建立在经验风险最小原则上的神经网络模型相比,理论上更为完善。本文运用支持向量机建立时间序列预测模型,研究影响模型预测精度的相关参数,在分析参数对时间序列预测精度的影响基础上,提出用遗传算法建立支持向量机预测模型的参数自适应优化算法。最后,用太阳黑子数据和航空发动机油样光谱数据进行了预测分析。算例表明了本文算法的正确性。

Support Vector Machine （SVM） is based on Statistical Learning Theory （SLT） and Structural Risk Minimization Principle （SRM）, and theoretically assures best generalization, therefore, it is theoretically better than Artificial Neural Network （ANN） which is based on Empirical Risk Minimization Principle （ERM）. In this paper, SVM was used to establish time series forecasting model, and on the basis of analyzing the influence of model parameters, a self-adaptive optimizing algorithm based on genetic algorithm was put forward. Finally, the sunspot data and the spectrometric oil data of some aero-engines were used for preliminary analysis, and the results show the correctness and validity of the new method.