Representing sentence structure in hidden Markov models for information extraction
Soumya Shubhra Ray, Mark W. Craven · 2001
We study the application of Hidden Markov Mod-els (HMMs) to learning information extractors for-ary relations from free text. We propose an ap-proach to representing the grammatical structure of sentences in the states of the model. We also in-vestigate using an objective function during HMM training which maximizes the ability of the learned models to identify the phrases of interest. We eval-uate our methods by deriving extractors for two bi-nary relations in biomedical domains. Our experi-ments indicate that our approach learns more accu-rate models than several baseline approaches. 1