基于BP神经网络优化的灰色线性回归组合模型应用分析

在分析灰色线性回归组合预测模型基本原理的基础上,利用MATLAB强大的计算功能,实现组合预测模型算法。通过实例分析发现拟合结果对实测值出现一定的波动性,故通过建立实测值与模拟值之间的比值序列,再利用BP神经网络模型对该比值序列进行建模优化,以进一步优化组合模型的预测精度。最后实例证明了该优化模型具有较高的拟合和预测精度,是一种可行、有效的优型变形数据分析模型。

Application Analysis of Based on Grey Linear Regression Combination Model Optimized by BP Neural Network

The article presents the algorithm to accomplish grey-linear regression model by using the MATLAB＇s powerful calculating function. However, the results of the example analysis show that there is volatility between the fitted values and measured values. The article establish the ratios be-tween measured values and analog values, then use BP neural network model for the ratio to opti-mized the prediction of combined model. Finally, the analyses of the example show that the optimized model has higher fitting and prediction accuracy and is a feasible and effective model for deformation data analysis.