基于熵权法的BDS钟差组合预测模型的建立

Establishment of BDS clock error combination prediction model based on entropy weight method

针对卫星钟差单一预测模型预测精度差,易造成较大偏差等问题,提出基于熵权法的钟差组合预测模型。该模型在对原始数据采用四分位数法和分段线性插值法完成数据预处理的基础上,分别采用二次多项式(QP)模型和灰色模型(GM)完成钟差预测,然后利用信息熵定权方法对单一模型赋权,进而得到QP-GM组合模型的预测结果。实验结果表明:组合预测模型在北斗卫星导航系统(BDS)钟差预测中的预测精度较单一模型均有提高,具有更好的适用性和可行性。

Aiming at the problems of poor prediction accuracy and large deviation caused by a single prediction model of satellite clock difference,a combined prediction model of clock difference based on entropy weight method is proposed.Based on the data preprocessing of the original data by quartile method and piecewise linear interpolation method,the Quadratic Polynomial(QP)model and Grey Model(GM)are used to complete the clock error prediction respectively.Then the information entropy weight method is used to weight the single model,and the prediction results of the QP-GM combined model are obtained.The first mock exam results show that the prediction accuracy of the combined forecasting model in BeiDou navigation satellite System(BDS)clock error prediction is higher than that of the single model,and has better applicability and feasibility.