נגישות
menu      
Advanced Search
Syntax
Search...
Volcani treasures
About
Terms of use
Manage
Community:
אסיף מאגר המחקר החקלאי
Powered by ClearMash Solutions Ltd -
Overcoming multicollinearity by deducting errors from the dependent variable
Year:
2001
Authors :
Schmilovitch, Ze'ev
;
.
Volume :
69
Co-Authors:
Pasternak, H., Institute of Agricultural Engineering, Volcani Center, A.R.O., Bet Dagan 50250, Israel
Edan, Y., Institute of Agricultural Engineering, Volcani Center, A.R.O., Bet Dagan 50250, Israel, Department of Industrial Engineering and Management, Ben-Gurion University of the Negev, Israel
Schmilovitch, Z., Institute of Agricultural Engineering, Volcani Center, A.R.O., Bet Dagan 50250, Israel
Facilitators :
From page:
761
To page:
768
(
Total pages:
8
)
Abstract:
A method to overcome multicollinearity, based on deduction of the stochastic disturbance from the dependent variable, was developed. This method resulted in a marked increase in the coefficient of determination - R2 associated with the "corrected dependent variable". Simulation results revealed that the proposed method yielded good estimation of those stochastic disturbance. The resulting "corrected" dependent variable may be incorporated into some of the existing methods for overcoming multicollinearity. Application of least-squares regression to the "corrected" dependent variable yields estimators of normal size, and the predictive ability of the derived equation is excellent. © 2001 Elsevier Science Ltd. All rights reserved.
Note:
Related Files :
computer simulation
Multicollinearity
regression analysis
Spectroscopic analysis
spectroscopy
Show More
Related Content
More details
DOI :
10.1016/S0022-4073(00)00111-4
Article number:
Affiliations:
Database:
Scopus
Publication Type:
article
;
.
Language:
English
Editors' remarks:
ID:
29035
Last updated date:
02/03/2022 17:27
Creation date:
17/04/2018 00:43
Scientific Publication
Overcoming multicollinearity by deducting errors from the dependent variable
69
Pasternak, H., Institute of Agricultural Engineering, Volcani Center, A.R.O., Bet Dagan 50250, Israel
Edan, Y., Institute of Agricultural Engineering, Volcani Center, A.R.O., Bet Dagan 50250, Israel, Department of Industrial Engineering and Management, Ben-Gurion University of the Negev, Israel
Schmilovitch, Z., Institute of Agricultural Engineering, Volcani Center, A.R.O., Bet Dagan 50250, Israel
Overcoming multicollinearity by deducting errors from the dependent variable
A method to overcome multicollinearity, based on deduction of the stochastic disturbance from the dependent variable, was developed. This method resulted in a marked increase in the coefficient of determination - R2 associated with the "corrected dependent variable". Simulation results revealed that the proposed method yielded good estimation of those stochastic disturbance. The resulting "corrected" dependent variable may be incorporated into some of the existing methods for overcoming multicollinearity. Application of least-squares regression to the "corrected" dependent variable yields estimators of normal size, and the predictive ability of the derived equation is excellent. © 2001 Elsevier Science Ltd. All rights reserved.
Scientific Publication
You may also be interested in