Comparing Two Regression Methods Based on Different Principal Components
-
摘要: 对两种采用成分提取进行回归建模的方法进行比较分析.指出采用主成分分析提取的主成分,虽然能很好地概括自变量系统中的信息,却往往对因变量缺乏解释能力.而采用PLS回归所提取的成分,则能在很好地概括自变量系统信息的同时,最好地解释因变量,排除无意义的信息干扰作用.Abstract: This paper deals with the comparing two regression methods based on different principal components.The principal components acquired by principal component analysis are the best summery of the information in the independent variables' system,but generally they are short of the explanatory capability to the dependent variable.However the components acquired by using PLS regression can not only summarize the independent variables well,but also they have the best explanation to the dependent variable.Moreover,the interference from invalid information is eliminated at the same time.
-
Key words:
- regression analysis /
- comparison /
- principal component analysis /
- PLS regression
-
1. Wold S,Albano C,Dunn III,et al.Pattern regression:Finding and using regularities in multivariate data.In:Martens J.Proc IUFOST Conf Food Research and Data.London:Analysis Applied Science Publication,1983 2. Tenenhaus M,Gauchi J P,Menardo C.Regression PLS et application.Revue de Statistiques Appliquees,1995,53(1):7~63 3. Marens H,Naes T.Multivariate calibration.New York:John Wiley & Sons,1989
点击查看大图
计量
- 文章访问数: 3288
- HTML全文浏览量: 80
- PDF下载量: 294
- 被引次数: 0