FASTUS: A System for Extracting Information from Natural-Language Text

摘要: Abstract : FASTUS is a system for extracting information from free text in English, and potentially other languages as well, entry into database, applications. It works essentially cascaded, nondeterministic finite state automaton. There are four steps the operation of FASTUS. In Step (1) sentences scanned certain trigger words to determine whether further processing should be done. (2) noun groups, verb prepositions some particles recognized. The input (3) sequence phrases recognized (2); patterns interest identified corresponding incident structures built up. (4) that derive same merged, these used generating database entries. an order magnitude faster than any comparable system; it can process news report average less eleven seconds. This translates directly fast development time. three half weeks between its first use MUC-4 evaluation May 1992, we were able build up domain knowledge point where was among leaders evaluation.