Abstract: The paper describes main features of formal lexico-syntactic pattern language (LSPL) proposed for specification of linguistics information about NL expressions automatically extracted from Russian texts. A fullyimplemented procedure for matching LSPL-patterns with text are presented, as well as developed programming tools for extraction of phrases specified by the patterns. Two applications of the language and the tools are discussed: terminological analysis of scientific texts and processing of NL sentences for question answering. LSPL-patterns developed for these applications are briefly characterized.

Keywords: information extraction from NL texts, lexico-syntactic patterns, matching procedure, automatic terms recognition and extraction, analysis of NL phrases for question answering.

ACM Classification Keywords: I.2.7 Artificial Intelligence: Natural language processing – Text analysis

Link:

LSPL-PATTERNS AS A TOOL FOR INFORMATION EXTRACTION FROM NATURAL LANGUAGE TEXTS

Elena Bolshakova, Natalia Efremova, Alexey Noskov

http://foibg.com/ibs_isc/ibs-16/ibs-16-p13.pdf