Application of CART and Minitab software to identify variables affecting indoor concentration levels

被引：21

作者：

Kadiyala, Akhil ^{[1
]}

Kumar, Ashok ^{[1
]}

机构：

[1] Univ Toledo, Dept Civil Engn, Toledo, OH 43606 USA

来源：

ENVIRONMENTAL PROGRESS | 2008年 / 27卷 / 02期

关键词：

D O I：

10.1002/ep.10292

中图分类号：

X [环境科学、安全科学];

学科分类号：

08 ; 0830 ;

摘要：

The use of two different softwares such as Minitab and CART involving regression analysis, for identifying the factors affecting the in-vehicle pollutant concentration, is proposed. Minitab regression analysis identifies the effect of two or more independent variables on the dependent pollutant concentration levels and thus identifies that passengers are an important variable affecting pollutant concentration for CO2. CART software develops a regression tree model and the CART algorithm identifies the important variables affecting the pollutant concentrations and ignores the insignificant variables. The results also indicate that CART does not require specification of any functional form as it is nonparametric and nonlinear, while Minitab provides linear relationship among the variables.

引用

页码：160 / 168

页数：9