似大地水准面的支持向量机模型研究  

Study on Support Vector Machine model for determination of quasi-geoid

作　　者：郝伟涛[1] 郭向前[1] 米川[1] 

机构地区：[1]河南省地质测绘总院,郑州450006

出　　处：《测绘科学》2012年第4期22-23,63,共3页Science of Surveying and Mapping

摘　　要：支持向量机（SVM）是一种基于结构风险最小化原理的学习技术,也是一种新的具有较好泛化性能的回归方法。本文简要介绍了SVM原理,针对大面积复杂似大地水准面的确定问题,仅依据测区的GPS水准实测数据,利用SVM方法整体建模。通过工程实例并与神经网络模型进行对比,证实了SVM似大地水准面模型的可靠性。ion： Support Vector Machine（SVM） is a learning technique based on the structural risk minimization principle,and it is also a class of regression method with good generalization ability.Aiming at the determination of large area complex quasi-geoid,only depending on the GPS leveling data,this paper first introduced the principle of SVM briefly,then chose the parameters and built the quasi-geoid model.Through taking an example and comparing with the Neural Network model,the correctness and effectiveness of the SVM model were demonstrated finally.

关 键 词：GPS水准 似大地水准面 支持向量机模型 

分 类 号：P223[天文地球—大地测量学与测量工程]