On the underestimation of model uncertainty by Bayesian K-nearest neighbors

Simple item page
dc.creatorSu, Wanhua
dc.creatorChipman, Hugh
dc.creatorZhu, Mu
dc.date2008-04-08
dc.date.accessioned2026-07-07T09:31:06Z
dc.date.available2026-07-07T09:31:06Z
dc.descriptionWhen using the K-nearest neighbors method, one often ignores uncertainty in the choice of K. To account for such uncertainty, Holmes and Adams (2002) proposed a Bayesian framework for K-nearest neighbors (KNN). Their Bayesian KNN (BKNN) approach uses a pseudo-likelihood function, and standard Markov chain Monte Carlo (MCMC) techniques to draw posterior samples. Holmes and Adams (2002) focused on the performance of BKNN in terms of misclassification error but did not assess its ability to quantify uncertainty. We present some evidence to show that BKNN still significantly underestimates model uncertainty.
dc.identifierhttps://arxiv.org/abs/0804.1325
dc.identifierhttp://arxiv.org/abs/0804.1325
dc.identifier.urihttp://salesiana.dossiersoluciones.com/handle/123456789/158341
dc.subjectMachine Learning
dc.titleOn the underestimation of model uncertainty by Bayesian K-nearest neighbors
dc.typetext

Files

Collections

arXiv