Fast bootstrap methodology for regression model selection Export

@article{citeulike:3885725, abstract = {Using resampling methods like cross-validation and bootstrap is a necessity in neural network design, for solving the problem of model structure selection. The bootstrap is a powerful method offering a low variance of the model generalization error estimate. Unfortunately, its computational load may be excessive when used to select among neural networks models of different structures or complexities. This paper presents the fast bootstrap (FB) methodology to select the best model structure; this methodology is applied here to regression tasks. The fast bootstrap assumes that the computationally expensive term estimated by the bootstrap, the optimism, is usually a smooth function (low-order polynomial) of the complexity parameter. Approximating the optimism term makes it possible to considerably reduce the necessary number of simulations. The FB methodology is illustrated on multi-layer perceptrons, radial-basis function networks and least-square support vector machines.}, author = {Lendassea, A. and Simonb, G. and Wertzc, V. and Verleysenb, M.}, citeulike-article-id = {3885725}, citeulike-linkout-0 = {http://dx.doi.org/10.1016/j.neucom.2004.11.017}, citeulike-linkout-1 = {http://linkinghub.elsevier.com/retrieve/pii/S092523120400517X}, doi = {10.1016/j.neucom.2004.11.017}, issn = {09252312}, journal = {Neurocomputing}, keywords = {ann, bootstrap, model-selection, resampling}, month = {March}, pages = {161--181}, posted-at = {2009-01-14 16:17:38}, priority = {2}, title = {Fast bootstrap methodology for regression model selection}, url = {http://dx.doi.org/10.1016/j.neucom.2004.11.017}, volume = {64}, year = {2005} }

TY - JOUR ID - citeulike:3885725 L3 - citeulike-article-id:3885725 N2 - Using resampling methods like cross-validation and bootstrap is a necessity in neural network design, for solving the problem of model structure selection. The bootstrap is a powerful method offering a low variance of the model generalization error estimate. Unfortunately, its computational load may be excessive when used to select among neural networks models of different structures or complexities. This paper presents the fast bootstrap (FB) methodology to select the best model structure; this methodology is applied here to regression tasks. The fast bootstrap assumes that the computationally expensive term estimated by the bootstrap, the optimism, is usually a smooth function (low-order polynomial) of the complexity parameter. Approximating the optimism term makes it possible to considerably reduce the necessary number of simulations. The FB methodology is illustrated on multi-layer perceptrons, radial-basis function networks and least-square support vector machines. JF - Neurocomputing SN - 09252312 EP - 181 TI - Fast bootstrap methodology for regression model selection VL - 64 SP - 161 KW - ann KW - bootstrap KW - model-selection KW - resampling AU - Lendassea, A AU - Simonb, G AU - Wertzc, V AU - Verleysenb, M PY - 2005/03// UR - http://dx.doi.org/10.1016/j.neucom.2004.11.017 ER -