Accurate Bayesian Data Classification without Hyperparameter Cross-validation
Methodology
2017-12-29 v1 Machine Learning
Statistics Theory
Statistics Theory
Abstract
We extend the standard Bayesian multivariate Gaussian generative data classifier by considering a generalization of the conjugate, normal-Wishart prior distribution and by deriving the hyperparameters analytically via evidence maximization. The behaviour of the optimal hyperparameters is explored in the high-dimensional data regime. The classification accuracy of the resulting generalized model is competitive with state-of-the art Bayesian discriminant analysis methods, but without the usual computational burden of cross-validation.
Keywords
Cite
@article{arxiv.1712.09813,
title = {Accurate Bayesian Data Classification without Hyperparameter Cross-validation},
author = {M Sheikh and A C C Coolen},
journal= {arXiv preprint arXiv:1712.09813},
year = {2017}
}