IBIMOL   23987
INSTITUTO DE BIOQUIMICA Y MEDICINA MOLECULAR PROFESOR ALBERTO BOVERIS
Unidad Ejecutora - UE
artículos
Título:
Replacement Method and Enhanced Replacement Method Versus the Genetic Algorithm Approach for the Selection of Molecular Descriptors in QSPR/QSAR Theories
Autor/es:
MERCADER, ANDREW G.; DUCHOWICZ, PABLO R.; FERNÁNDEZ, FRANCISCO M.; CASTRO, EDUARDO A.
Revista:
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES
Editorial:
ACS Publications
Referencias:
Año: 2010 vol. 50 p. 1542 - 1548
ISSN:
0095-2338
Resumen:
Linear methods are usually considered as the best screening techniques for analyzing QSPR/QSAR datasets in order to identify the most relevant structural features. In present work,  we compared extensively our recently developed Enhanced Replacement Method (ERM) and Replacement Method against the Genetic Algorithm (GA) approach for the selection of an optimal subset of molecular descriptors from a much greater pool of such regression variables. Our approaches yielded optimal results with a much smaller number of linear regressions than the exact (full) combinatorial search of descriptors. The methods were tested on seven different experimental datasets. The comparison showed that ERM is preferable than GA. On the other hand, GA gave similar results to RM; nevertheless the latter is much simpler and easier to use in practice.