主成分分析在GPS高程拟合中的应用  

Application of principal component analysis in GPS height fitting

作　　者：岳东杰[1] 李立瑞[2] 郭培闪[1] 

机构地区：[1]河海大学地球科学与工程学院,江苏南京210098 [2]中国电力工程顾问集团中南电力设计院,湖北武汉430071

出　　处：《河海大学学报（自然科学版）》2011年第3期323-326,共4页Journal of Hohai University(Natural Sciences)

基　　金：国家"十一五"科技支撑计划(2006BAG04B04)

摘　　要：为了合理确定多项式的阶数,克服参数间的相关性,提高GPS高程拟合的精度,提出采用主成分分析法,对相同的数据点分别采用常规多项式拟合与主成分分析拟合.苏通大桥首级施工控制网的GPS高程拟合计算结果表明,主成分分析法受参数个数与相关性影响小,计算结果稳定,而且能有效降低数据冗余度,是一种比较好的拟合方法.In order to rationally determine the order of polynomial,to overcome the correlativity among parameters and to raise the fitting precision of GPS height,it was put forward that the conventional polynomial fitting and the principal component analysis were respectively employed for the same data point by means of the principal component analysis method.The calculated results of GPS height fitting for the first stage construction control network of Su-Tong Bridge indicate that the principal component analysis method is little affected by the number and the correlativity of parameters and has stable computational results.It can effectively reduce the data redundancy,thus it is a good method for fitting.

关 键 词：多项式拟合 GPS高程拟合 主成分分析法 苏通大桥 

分 类 号：TU198.2[建筑科学—建筑理论]