TY  - JOUR
T1  - Mathematical Programming Approach for Variable Selection in Discriminant
Analysis: An Environmental Application
AU - Majid Nassar, Aiman AU - Ahmed Saleh, Anwer 
JO  - Journal of Engineering and Applied Sciences
VL  - 14
IS  - 8
SP  - 2477
EP  - 2484
PY  - 2019
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2019.2477.2484
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2019.2477.2484
KW  - Discriminant analysis
KW  -mixed integer programming
KW  -variable selection
KW  -mathematical
programming
KW  -environmental impact assessment
KW  -environmental
AB  - The environmental impact assessment of projects is based on several variables. It is desirable to
select the most important variables needed for classifying any project in only one of three classes according
to severity of possible environmental impact. In this study, the selection of variables in discriminant analysis
between more than two groups using the Mathematical programming approach is applied to Egypt
environmental impact assessment survey 2000 data to select the most important variables for classifying new
projects in the true classes. The results are compared with those of the stepwise method. The comparison
shows that according to the percent of correct classification, the Mathematical programming model is better
than the stepwise method.
ER  -