H
Harry Zhang
Researcher at University of New Brunswick
Publications - 73
Citations - 5231
Harry Zhang is an academic researcher from University of New Brunswick. The author has contributed to research in topics: Naive Bayes classifier & Bayesian programming. The author has an hindex of 28, co-authored 73 publications receiving 4654 citations. Previous affiliations of Harry Zhang include Eaton Corporation.
Papers
More filters
Proceedings Article
The Optimality of Naive Bayes.
TL;DR: A sufficient condition for the optimality of naive Bayes is presented and proved, in which the dependence between attributes do exist, and evidence that dependence among attributes may cancel out each other is provided.
Proceedings Article
AUC: a statistically consistent and more discriminating measure than accuracy
TL;DR: It is formally proved that, for the first time, AUC is a better measure than accuracy in the evaluation of learning algorithms.
Book ChapterDOI
AUC: a better measure than accuracy in comparing learning algorithms
TL;DR: R rigourously is established that, even in this setting, the area under the ROC (Receiver Operating Characteristics) curve, or simply AUC, provides a better measure than accuracy when measuring and comparing classification systems.
Journal ArticleDOI
A Novel Bayes Model: Hidden Naive Bayes
TL;DR: This paper summarizes the existing improved algorithms and proposes a novel Bayes model: hidden naive Bayes (HNB), which significantly outperforms NB, SBC, NBTree, TAN, and AODE in terms of CLL and AUC.
Journal ArticleDOI
Exploring conditions for the optimality of naïve bayes
TL;DR: This paper designs a dependence distribution-based algorithm by extending the ChowLiu algorithm, a widely used CI based algorithm, and presents and proves a sufficient condition for the optimality of Naive Bayes, in which the dependences among attributes exist.