On Discriminative vs. Generative classi ers: A comparison of logistic regression and naive Bayes

by: Andrew Y Ng, Michael I Jordan

(2001)

View FullText article

CiteSeerX Alpha

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

vlachmore

Find related articles with these CiteULike tags

classification, discriminative, generative

摘要

We compare discriminative and generative learning as typied by logistic regression and naive Bayes. We show, contrary to a widelyheld belief that discriminative classiers are almost always to be preferred, that there can often be two distinct regimes of performance as the training set size is increased, one in which each algorithm does better. This stems from the observation|which is borne out in repeated experiments|that while discriminative learning has lower asymptotic error, a generative classier may also approach its (higher) asymptotic error much faster. 1

@misc{citeulike:3329371, abstract = {We compare discriminative and generative learning as typied by logistic regression and naive Bayes. We show, contrary to a widelyheld belief that discriminative classiers are almost always to be preferred, that there can often be two distinct regimes of performance as the training set size is increased, one in which each algorithm does better. This stems from the observation|which is borne out in repeated experiments|that while discriminative learning has lower asymptotic error, a generative classier may also approach its (higher) asymptotic error much faster. 1}, author = {Ng, Andrew Y. and Jordan, Michael I. }, citeulike-article-id = {3329371}, keywords = {classification, discriminative, generative}, posted-at = {2008-09-24 14:35:11}, priority = {2}, title = {On Discriminative vs. Generative classi ers: A comparison of logistic regression and naive Bayes}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.19.9829}, year = {2001} }

RIS record

TY - ELEC ID - citeulike:3329371 L3 - citeulike-article-id:3329371 N2 - We compare discriminative and generative learning as typied by logistic regression and naive Bayes. We show, contrary to a widelyheld belief that discriminative classiers are almost always to be preferred, that there can often be two distinct regimes of performance as the training set size is increased, one in which each algorithm does better. This stems from the observation|which is borne out in repeated experiments|that while discriminative learning has lower asymptotic error, a generative classier may also approach its (higher) asymptotic error much faster. 1 TI - On Discriminative vs. Generative classi ers: A comparison of logistic regression and naive Bayes KW - classification KW - discriminative KW - generative AU - Ng, Andrew AU - Jordan, Michael PY - 2001/// UR - http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.19.9829 ER -

RIS BibTeX