Information Extraction Based on Semantic Tagging
LI Xiang-yang, Pla Uni · Journal of PLA University of Science and Technology · 2004
As part of speech tagging is the main reason for semantic lose, semantic tagging is suggested to be the base for rule-building in information extraction. Extraction based on semantic tagging can eliminate three negative effects caused by part of speech tagging and uses a uniform approach to guide the process of information extraction. Processing at a finer granularity level, the method has a universal appeal for information extraction in domains showing strong semantic rules. A system called MIE (a military information extraction system), which is based on semantic tagging, is designed and used to extract information from scenario texts. The result shows that the approach is feasible and has practical promise.