神经网络方法在地表高程预测中的应用研究

Applied Research on Neural Network in Surface Elevation Prediction

从地表高程预测的研究现状出发,基于Matlab神经网络工具箱构建了广义回归神经网络(GRNN),建立了能够预测地表高程的神经网络预测模型,并与反距离加权法(IDW)的预测结果进行了比较分析。结果表明,GRNN方法计算简单,在相同的样本数据条件下,GRNN的预测精度较高,表明神经网络方法在地表高程预测中的应用是可行的。

This paper proceeded from the actuality of surface elevation prediction research,structured General Regression Neural Network（GRNN）based on Matlab artificial neural network toolbox,and established the neural network prediction model which could predict surface elevation.The predicted results between GRNN and Inverse Distance Weighting（IDW） were compared simultaneously.Results showed that GRNN could get the better prediction results with a simple algorithm under the same sample condition.Therefore,GRNN could be used to predict surface elevation.