摘要
为了提高区域GNSS高程拟合模型的预测精度,针对最小二乘支持向量机(LSSVM)模型中难以选择最佳参数的问题,将鲸鱼算法(WOA)引入最小二乘支持向量机中,利用其全局搜索能力强、参数少、收敛速度快等特性,为常规最小二乘支持向量机提供可靠的核参数和正则化参数。采用少量实际工程中的GNSS水准重合点进行检验,选择高程异常值的残差均方根误差作为组合算法建模精度的评判依据。结果表明:在带状区域中,WOA-LSSVM拟合模型的外符合精度相对于常规LSSVM拟合模型提高了30.3%;在小范围面状区域中,WOA-LSSVM拟合模型的精度、稳定性也优于LSSVM、BP拟合法,为今后GNSS高程拟合模型的建立提供了一定的参考价值。
In order to improve the prediction accuracy of the regional GNSS elevation fitting model,it is difficult to select the best parameters in the LSSVM model.WOA is introduced into the least squares support vector machine.With strong global search capability,few parameters,and fast convergence speed,it provides reliable kernel parameters and regularization parameters for conventional least squares support vector machines.Then a small number of GNSS level coincidence points is used in the actual project in the test,and residual root mean square error of the elevation outliers is selected as the evaluation basis of the combined algorithm modeling accuracy.The results show that the external coincidence accuracy of the WOA-LSSVM fitting model in the band area is 30.3%,higher than that of the conventional LSSVM fitting model.In a small area,the accuracy and stability of the WOA-LSSVM fitting model are also better than the LSSVM and BP fitting methods,providing reference for the establishment of GNSS elevation fitting model in the future.
作者
何广焕
唐诗华
邢鹏威
张跃
蒙金龙
HE Guang-huan;TANG Shi-hua;XING Peng-wei;ZHANG Yue;MENG Jin-long(College of Geomatics and Geoinformation,Guilin University of Technology,Guilin 541006,China;Guangxi Key Laboratory of Spatial Information and Geomatics,Guilin University of Technology,Guilin 541006,China)
出处
《桂林理工大学学报》
CAS
北大核心
2021年第4期837-842,共6页
Journal of Guilin University of Technology
基金
国家自然科学基金项目(41864002)
广西自然科学基金项目(2018GXNSFAA281279)
广西高校中青年教师科研基础能力提升项目(KY2016YB823)。
关键词
高程拟合
鲸鱼算法
最小二乘支持向量机
正则化参数
核参数
elevation fitting
whale optimization algorithm(WOA)
least square support vector machine(LSSVM)
regularization parameters
kernel parameters
