摘要
将二次曲面、BP神经网络、最小二乘支持向量机应用与高程异常拟合,并用某地区数据进行了实验验证,结果表明,最小二乘支持向量机应用于高程异常拟合精度最优。
The quadraftc surface model BP neural network model and least squares support vector machine （LS-SVM） model is applied to the GPS height anomaly fitting. The GPS elevation data in a certain area is used, the results shows LS- SVM model would be significantly better than quadratic surface model and BP neural network model.
出处
《北京测绘》
2014年第6期20-22,共3页
Beijing Surveying and Mapping
关键词
高程异常
二次曲面
BP神经网络
最小二乘支持向量机
GPS height anomaly fitting
quadraftc surface model
BP neural network models least squares support vector
