Extraction of Chemical Reaction Information from Primary Journal Text Using Computational Linguistics Techniques. 1. Lexical and Syntactic Phases

Abstract

An experimental technique for extracting specific facts about chemical reactions from the full text of American Chemical Society journals is described. The chemical information is used to encode Reaction Information Forms that can serve as the basis for a computer-readable file of chemical reactions. The programs use computational linguistics techniques to analyze the natural language text. The restricted semantic domain of the source publications makes it possible to resolve ambiguities effectively. This paper focuses on the lexical and syntactic aspects of the specialized information extraction procedures, and the following paper will describe the associated semantic considerations. © 1984, American Chemical Society. All rights reserved.