灰色最小支持向量机模型在高程拟合中的应用

Application of grey least square support vector machine model in height fitting

针对传统单一灰色最小二乘支持向量机(GLSSVM)高程拟合方法的不足以及LSSVM模型参数选择的随机性,该文提出了一种基于PSO-GA算法优化的灰色最小二乘支持向量机高程拟合模型。模型将灰色模型与最小二乘支持向量机模型相结合,建立GLSSVM模型,并结合粒子群算法与遗传优化算法寻找GLSSVM模型的最优参数组合。为进一步验证提出模型的可靠性与有效性,通过具体工程实例,并将拟合结果分别与粒子群算法优化的最小二乘支持向量机模型(PSO-GLSSVM),遗传算法优化的最小二乘支持向量机模型(GA-GLSSVM)及单一GLSSVM模型进行对比分析,结果表明,PSO-GA-GLSSVM模型拟合精度更好,可靠性更高,为高程拟合研究提供了一种思路。

Aiming at the shortcomings of traditional single LSSVM elevation fitting method and the randomness of least squares support vector machine model parameter selection,a grey least squares support vector machine elevation fitting model based on PSO-GA optimization algorithm was proposed.The grey model and least squares support vector machine model was combined,then the GLSSVM was established.Particle swarm optimization algorithm and genetic algorithm was adapted to find the optimal parameter combination of GLSSVM model,In order to verify the reliability and efficiency of the model,practical examples and the PSO-GLSSVM,GA-G LSSVM,GLSSVM model were compared with the fitting result.The results showed that the new model had better accuracy and better reliability and provide an idea for elevation fitting studies.