Automatic Extraction of Analytical Chemical Information. System Description, Inventory of Tasks and Problems, and Preliminary Results

A system for semiautomatic extraction of information from abstracts describing analytical methods is described. The system is based on the theory of Government and Binding for the syntactic part and Conceptual Graphs for the discourse analysis part. The system is modular and largely domain independent. The corpus of abstracts for which it is being developed contains abstracts from Analytical Abstracts Online. The current status of the system is that the grammar is finished and that the lexicon and the discourse module are under development. Preliminary results are that the system is capable of analyzing six abstracts on various analytical techniques.