基于广义回归神经网络的路基沉降预测

为了对路基沉降变化规律进行预测，避免发生工程事故，提出了将广义回归神经网络模型应用于软土地基沉降预测中的方案。通过广义回归神经网络的基本理论和概念，采用实际工程数据，用 BP 神经网络方法和广义回归神经网络方法进行了预测分析，比较了2种方法的3组预测结果。工程实例预测结果表明，广义回归神经网络方法的均方误差和决定系数表现都优于 BP 神经网络方法；证明该方法是可行且有效的。

Subgrade Settlement Prediction Based on Support Vector Machine for Regression

In order to predict the variation of subgrade settlement,to avoid projects accidents,general-ized regression neural network model in subgrade settlement prediction program is proposed.Firstly, the basic theories and concepts of generalized regression neural network are discussed;Secondly, based on the actual project data,generalized regression neural network method and BP neural network model are applied for prediction and analysis;Finally,three groups based on two prediction methods are compared.Projects prediction results show that mean square error and the coefficient of determi-nation performance of generalized regression neural network model are better than BP neural network＇s;and the results prove that the method is feasible and effective.