| BP神经网络和SVR用于GPS高程拟合研究分析 |
| |
| 引用本文： | 陈德忠,赵亮,何书静,叶世榕.BP神经网络和SVR用于GPS高程拟合研究分析[J].测绘信息与工程,2012,37(5):47-49. |
| |
| 作者姓名： | 陈德忠 赵亮 何书静 叶世榕 |
| |
| 作者单位： | 1. 武汉大学卫星导航定位技术研究中心,武汉市珞喻路129号,430079
2. 福建省测绘院,福州市华林路205号,350003 |
| |
| 基金项目： | 国家自然科学基金资助项目（41074008） |
| |
| 摘 要： | 首先介绍BP神经网络和SVR方法(支持向量机回归)用于GPS高程拟合的原理,然后通过实际数据比较BP算法和SVR在GPS高程拟合中精度。结果表明,以结构风险最小化为准则的学习方法SVR,其泛化能力明显比BP神经网络好,在工程中具有一定的实际应用价值。
|
| 关 键 词： | BP神经网络 SVR GPS高程拟合 |
Analysis of GPS Elevation Fitting Based on BP Neural Network and SVR |
| |
| CHEN Dezhong,ZHAO Liang,HE Shujing,YE shirong.Analysis of GPS Elevation Fitting Based on BP Neural Network and SVR[J].Journal of Geomatics,2012,37(5):47-49. |
| |
| Authors: | CHEN Dezhong ZHAO Liang HE Shujing YE shirong |
| |
| Institution: | 1 Research Center of GNSS,Wuhan University,129 Luoyu Road,Wuhan 430079,China; 2 Surveying and Mapping Institute of Fujian Province,205 Hualin Road,Fuzhou 350003,China) |
| |
| Abstract: | First we introduce the theory of BP neural network and SVR（support vector machine regression） for GPS elevation Fitting.Then that compare the precision betwen BP algorithm and SVR in GPS elevation fitting with the case data.The results show learning method of SVR based on the structural risk minimization criterion is significantly better than the BP neural network on generalization ability,also has some real value in our engineering application. |
| |
| Keywords: | BP neural network SVR GPS elevation Fitting |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
