CiteSeerX — Naive Bayesian Learning

Naive Bayesian Learning (1997) [52 citations — 2 self]

Popular Tags

No tags have been applied to this document.

BibTeX | Add To MetaCart

@TECHREPORT{Elkan97naivebayesian,
    author = {Charles Elkan},
    title = {Naive Bayesian Learning},
    institution = {},
    year = {1997}
}

Years of Citing Articles

Bookmarks

OpenURL

Abstract:

Introduction So-called "naive" Bayesian classification is the optimal method of supervised learning if the values of the attributes of an example are independent given the class of the example. Although this assumption is almost always violated in practice, recent work has shown that naive Bayesian learning is remarkably effective in practice and difficult to improve upon systematically [Domingos and Pazzani, 1996] . On many real-world example datasets naive Bayesian learning gives better test set accuracy than any other known method, including backpropagation and C4.5 decision trees. Also, these classifiers can be learned very efficiently. Given e training examples over f attributes, the time required to learn a boosted naive Bayesian classifier is O(ef), i.e. linear. No learning algorithm that examines all its training data can be faster. 2 A review of naive Bayesian learning Let A

Citations

392	Perceptrons: an introduction computational geomery – Minsky, Papert - 1969
304	Supervised and unsupervised discretization of continuous features – Dougherty, Kohavi, et al. - 1995
234	Beyond independence: Conditions for the optimality of the simple Bayesian classifier – Domingos, Pazzani - 1996