基于支持向量机的不等时距灰色组合预测模型

建立了一种基于支持向量机的不等时距灰色组合预测模型.利用各种不等时距灰色模型的预测结果作为支持向量机的输入,实测值作为支持向量机的输出,并采用LS-SVM回归算法和高斯核函数对支持向量机进行训练,利用训练好的支持向量机即可进行组合预测.该模型兼具灰色模型所需原始数据少、建模简单、运算方便的优势和最小二乘支持向量机具有泛化能力强、非线性拟合性好、小样本等特性,弥补了单一不等时距预测模型的不足,避免了神经网络组合预测易于陷入局部最优的弱点.模型结构简单、实用,仿真结果验证了其有效性.

Forecast Model for Unequal Interval Grey Combination Based on the Support Vector Machine

A forecast model for grey combination based on the support vector machine（SVM） is presented in the paper.The input of SVM is the forecasting value of various grey model of unequal Interval and the output of SVM is actual value.SVM trained with LS-SVM regression algorithm and Gauss kernel function has the ability to give combination forecast.The model has combined the advantages of grey model such as less raw data to be required,simple to model,convenient to calculate,and features of LS-SVM such as strong generalization ability,good nonlinear fitting ability and less samples to be required,meanwhile it remedies the defect in single unequal Interval model,avoids the defect in neural networks combination forecast to fall easily into its local optima.The model construction is simple and practical,then validity of the method is proved with simulation results.