Editing Markov random field (section)

== Conditional random fields ==
{{Main|Conditional random field}}
One notable variant of a Markov random field is a '''[[conditional random field]]''', in which each random variable may also be conditioned upon a set of global observations <math>o</math>. In this model, each function <math>\varphi_k</math> is a mapping from all assignments to both the [[Clique (graph theory)|clique]] ''k'' and the observations <math>o</math> to the nonnegative real numbers. This form of the Markov network may be more appropriate for producing [[discriminative model|discriminative classifiers]], which do not model the distribution over the observations. CRFs were proposed by [[John D. Lafferty]], [[Andrew McCallum]] and [[Fernando C.N. Pereira]] in 2001.<ref name=ICML03classic>{{cite web |url=http://icml.cc/2013/?page_id=21 |title=Two classic paper prizes for papers that appeared at ICML 2013 |date=2013 |website=ICML |access-date=15 December 2014}}</ref>