An Optimal Self‐Pruning Neural Network and Nonlinear Descriptor Selection in QSAR
- 16 October 2009
- journal article
- research article
- Published by Wiley in QSAR & Combinatorial Science
- Vol. 28 (10) , 1092-1097
Abstract
Feature selection is an important but still poorly solved problem in QSAR modeling. We employ a Bayesian regularized neural network with a sparse Laplacian prior as an efficient method for supervised feature selection, and robust parsimonious nonlinear QSAR modeling. The method simultaneously selects the most relevant descriptors for model, and automatically prunes the neural network to have the architecture with optimum prediction ability. We illustrate the advantages of the method using a suite of diverse data sets, and compare the results obtained by the new method against those obtained by alternative contemporary methods.Keywords
This publication has 24 references indexed in Scilit:
- Toward Novel Universal Descriptors: Charge FingerprintsJournal of Chemical Information and Modeling, 2009
- Modelling blood–brain barrier partitioning using Bayesian neural netsJournal of Molecular Graphics and Modelling, 2004
- Nonlinear Multivariate Regression Outperforms Several Concisely Designed Neural Networks on Three QSPR Data SetsJournal of Chemical Information and Computer Sciences, 2000
- Atomistic Topological Indices Applied to Benzodiazepines using Various Regression MethodsQuantitative Structure-Activity Relationships, 1998
- A Chemically Intuitive Molecular Index Based on the Eigenvalues of a Modified Adjacency MatrixQuantitative Structure-Activity Relationships, 1997
- Bayesian Learning for Neural Networks1996
- Bayesian Regularization and Pruning Using a Laplace PriorNeural Computation, 1995
- A Practical Bayesian Framework for Backpropagation NetworksNeural Computation, 1992
- Synthesis and muscarinic activities of quinuclidin-3-yltriazole and -tetrazole derivativesJournal of Medicinal Chemistry, 1992
- Comparison of azabicyclic esters and oxadiazoles as ligands for the muscarinic receptorJournal of Medicinal Chemistry, 1991