- A CRF model consists of
- F = <f1,…,fk>, a vector of "feature functions"
- θ=<θ1,…,θk>, a vector of weights for each feature function
- Let O = <o1,…,oT> be an observed sequence
- Let A = <a1,…,aT> be the latent variables
p(A=y|O)=exp(θ⋅F(y,O))∑y′exp(θ⋅F(y′,O))
Reference
http://knight.cis.temple.edu/~yates/cis8538/sp11/slides/conditional-random-fields.ppt