摘要
采用最小二乘拟合求取概率积分法预计参数时,经常遇到发散问题,致使求参结果失真。而求参过程中法矩阵的病态是造成结果发散的主要原因之一。将岭估计应用于预计参数的求取中,能够有效地克服传统最小二乘拟合求参时法矩阵的病态。在给出基于岭估计求参模型的同时,使用Matlab实现了参数的自动求取。最后通过模拟算例验证了该求参模型克服法矩阵病态的有效性及求参结果的可靠性。
Divergence problem will often result in parameters distortion in solve prediction parameters of probability integration method with least square method. Morbidity of normal matrix is one of main divergence reasons. Ridge estimate was applied in solving prediction parameters, which could effectively overcome distortion of normal matrix. At the same time, Matlab was used to realize parameters" automatic solving. A simulation example was present to verify the model.
出处
《煤矿开采》
北大核心
2012年第2期17-19,85,共4页
Coal Mining Technology
基金
国家自然科学基金重点项目(50834004)
国家公益性行业基金资助项目(200811050)
中国矿业大学青年科研基金资助项目(09091252)
关键词
岭估计
概率积分法
求参
开采沉陷
L曲线法
ridge estimate
probability integration method
parameter solve
mining subsidence
L curve method