针对径向基神经网络(RBFNN)预报模型的不足,提出了一种基于马尔科夫链修正的RBFNN预报模型,以RBFNN模型的预测结果为基准,利用马尔科夫链进行误差修正,进而显著提高模型的预报精度。以某碾压混凝土重力坝的变形监测为例,建立大坝变形预报模型,并将其结果与单一的RBFNN模型的预报结果做了对比,结果表明,基于马尔科夫链修正的RBFNN预报模型精度更高,结果更符合实际。 Aiming at the deficiency of RBFNN forecasting model, this paper proposes a RBFNN forecasting model based on Markov Chain Correction. Based on the prediction results of RBFNN model, the Markov chain error correction is used to improve the RBFNN forecasting model. Model prediction accuracy. Taking deformation monitoring of a certain RCC gravity dam as an example, a dam deformation forecasting model is established, and its result is compared with that of a single RBFNN model. The results show that the accuracy of RBFNN forecasting model based on Markov Chain Correction Higher, the result is more realistic.