CAJ | 학술논문

研究混沌粒子群支持向量机在GPS高程拟合中的应用,考虑地形起伏对高程转换的影响,引入地形改正量构建新的支持向量机训练模型,并针对支持向量机的参数人为选择的盲目性,将混沌粒子群优化理论用于SVM参数的选取,并与传统的拟合算法如二次曲面法、多面函数法和BP神经网络法的比较结果表明其精度更优.
연구혼돈입자군지지향량궤재GPS고정의합중적응용,고필지형기복대고정전환적영향,인입지형개정량구건신적지지향량궤훈련모형,병침대지지향량궤적삼수인위선택적맹목성,장혼돈입자군우화이론용우SVM삼수적선취,병여전통적의합산법여이차곡면법、다면함수법화BP신경망락법적비교결과표명기정도경우.
The application of chaos particle swarm support vector machine in GPS height fitting is studied. Taking the impact of terrain on the height conversion into account, the terrain correction is introduced to the support vector machine model. Considering the blindness of man-made choice of parameters of SVM, a chaos particle swarm optimization theory is used to select the parameters of SVM. Compared with the traditional fitting methods, such as polynomial curved surface, multi-face function BP neural network, this method is better.