无用信息变量去除方法在变量选择过程中的应用

Method of removing unwanted information variables in the variable selection process application

本文研究了使用主成分回归及偏最小二乘法(pls)对煤炭热解产物红外光谱数据进行变量优选,剔除无用信息变量,建立了烟气成分与光谱数据关联的数学模型,并对多组实际数据进行分析,得到了满意的结果。

This paper studies the use of principal component regression and partial least squares(pls) of coal pyrolysis products preferred infrared spectral data variables,excluding variable useless information,the establishment of the flue gas composition and spectral data associated mathematical model,and multigroup to analyze actual data obtained satisfactory results