摘要
基于正则化理论,通过添加1个正则化因子,结合多项式回归模型,本文构造了正则化的钢管混凝土轴压短柱极限承载力分析模型。证明了存在正则因子,使多项式回归模型中信息矩阵的条件数可以降低到一定值,给出了计算公式。大量试验数据分析表明,新的正则化钢管混凝土轴压短柱极限承载力分析模型可改善多项式回归模型中信息矩阵的病态程度,避免由于信息矩阵的严重病态而导致矩阵求逆计算的失败,且正则化的分析模型与试验数据具有更小的拟合误差。
Based on regularization theory, by adding a regularization factor we constructed an analysis model for regularized ultimate bearing capacity column. It is proved that the existence of regular factor which can reduce the to polynomial regression model, of axial compressed CFST stub condition number of information matrix in polynomial regression model to a certain value, and the calculation formulas are also given. A large number of experimental data analysis shows that the new regularization theory analysis model can reduce the ill-conditioned extent to the information matrix in polynomial regression model, and avoid the inversion computing failure caused by severely ill-conditioned problem in information matrix, and the regularization theory analysis model has a better goodness of fitting than polynomial regression model.
出处
《公路交通科技》
CAS
CSCD
北大核心
2014年第2期81-85,共5页
Journal of Highway and Transportation Research and Development
基金
道路结构与材料交通重点实验室开放基金项目(kfj100209)
中国博士后科学基金面上项目(2013M531776)
关键词
桥梁工程
极限承载力
正则化理论
钢管混凝土轴压短柱
病态矩阵
bridge engineering
ultimate bearing capacity
regularizadon theory
axial compressed concrete-filled steel tubular (CFST) stub column
ill-conditioned matrix