๐”– Bobbio Scriptorium
โœฆ   LIBER   โœฆ

Comparison of Forward Selection, Backward Elimination, and Generalized Simulated Annealing for Variable Selection

โœ Scribed by J.M. Sutter; J.H. Kalivas


Publisher
Elsevier Science
Year
1993
Tongue
English
Weight
360 KB
Volume
47
Category
Article
ISSN
0026-265X

No coin nor oath required. For personal study only.

โœฆ Synopsis


In order to obtain quantitative information, it is often necessary for a chemist to employ regression methods. An algorithm is described for determining the optimal subset of variables which gives the best prediction in regression analysis. The procedure is based on the generalized simulated annealing method (GSA) of optimization. From a comparison study with standard methods of variable subset selection by forward selection and backward elimination, GSA is found to perform better. Three data sets are used for distinction purposes. Two of the data sets are relatively small, allowing comparison to the global variable subset obtained by computing all possible variable combinations. e 1993 Academic Press. Inc.


๐Ÿ“œ SIMILAR VOLUMES


Comparison of centroid and simulation ap
โœ John Butler; David L. Olson ๐Ÿ“‚ Article ๐Ÿ“… 1999 ๐Ÿ› John Wiley and Sons ๐ŸŒ English โš– 130 KB

the Department of Energy, Office of Fissile Materials Disposition completed a comparative analysis of alternative methods to dispose weapons-grade plutonium and made a final recommendation. This analysis considered the three overall objectives of minimizing proliferation of plutonium, selecting the