摘要
采用L曲线法大致定位迭代初始点,结合牛顿法加速求解最佳光滑系数.数值试验表明,经验值可能造成不收敛,且定位不够准确,而L曲线法直观、有效地定位初始点,保证了收敛性.通过GCV牛顿法提高了求解速度,在精度和时间上都取得较好效果.
To positione initial point by L-curve method,and solve smoothing parameter through Newton iterative method.Numerical experiments show that initial point with experience may do not converge,or not accurate enough.While intuitive L-curve method efficiently located the initial point,ensuring the convergence.GCV-Newton iterative method improved calculation speed,with a better accuracy and efficiency.
出处
《太原师范学院学报（自然科学版）》
2014年第2期4-9,共6页
Journal of Taiyuan Normal University:Natural Science Edition
关键词
函数型数据
光滑系数
广义交叉验证
L曲线法
牛顿法
Functional data
Smooth parameter
Generalized cross validation（GCV）
L curve method
Newton iterative method
