摘要
灰色模型具有计算简便、所需数据量小等优点,在短期预测中有较高的精度,但对波动性较大的数据预测效果较差;神经网络具有较强的非线性映射能力以及自学习能力,对波动性数据的处理效果良好。本文结合二者的优点,建立了GM-BP组合模型。实例证明,组合模型相对单一的模型在地表变形预测中具有明显的优势。
Gray model has advantages of calculation, required data and high precision in the short - term forecasting, but it impactsuncertain data negatively. Neural networksmodelforecasts uncertain data effectively by the ability of nonlinear mapping and self - learn-ing. Then the GM - BP combined model is purposed to integrate advantages of both. The calculation results show that the GM - BPcombined model has obvious advantages in ground deformation forecasting in contrast toSingle model.
出处
《测绘与空间地理信息》
2017年第4期195-197,共3页
Geomatics & Spatial Information Technology
